• Back, L. E., , and C. S. Bretherton, 2005: The relationship between wind speed and precipitation in the Pacific ITCZ. J. Climate, 18 , 43174328.

    • Search Google Scholar
    • Export Citation
  • Bretherton, C. S., , T. Uttal, , C. W. Fairall, , S. Yuter, , R. Weller, , D. Baumgardner, , K. Comstock, , and R. Wood, 2004: The EPIC 2001 stratocumulus study. Bull. Amer. Meteor. Soc., 85 , 967977.

    • Search Google Scholar
    • Export Citation
  • Cronin, M. F., , N. A. Bond, , C. Fairall, , J. Hare, , M. J. McPhaden, , and R. A. Weller, 2002: Enhanced oceanic and atmospheric monitoring underway in eastern Pacific. Eos, Trans. Amer. Geophys. Union, 83 , 210211.

    • Search Google Scholar
    • Export Citation
  • de Szoeke, S. P., , and C. S. Bretherton, 2004: Quasi-Lagrangian large eddy simulations of cross-equatorial flow in the east Pacific atmospheric boundary layer. J. Atmos. Sci., 61 , 18371858.

    • Search Google Scholar
    • Export Citation
  • de Szoeke, S. P., , C. S. Bretherton, , N. A. Bond, , M. F. Cronin, , and B. M. Morley, 2005: EPIC 95°W observations of the eastern Pacific atmospheric boundary layer from the cold tongue to the ITCZ. J. Atmos. Sci., 62 , 426442.

    • Search Google Scholar
    • Export Citation
  • Hartten, L. M., , and K. S. Gage, 2000: ENSO’s impact on the annual cycle: The view from Galapagos. Geophys. Res. Lett., 27 , 385388.

  • Huffman, G. J., and Coauthors, 2001: Global precipitation at one-degree daily resolution from multisatellite observations. J. Hydrometeor., 2 , 3650.

    • Search Google Scholar
    • Export Citation
  • Lindzen, R. S., , and S. Nigam, 1987: On the role of sea surface temperature gradients in forcing low-level winds and convergence in the tropics. J. Atmos. Sci., 44 , 24182436.

    • Search Google Scholar
    • Export Citation
  • Madden, R. A., , and P. R. Julian, 1971: Detection of a 40–50 day oscillation in the zonal wind in the tropical Pacific. J. Atmos. Sci., 28 , 702708.

    • Search Google Scholar
    • Export Citation
  • Maloney, E. D., , and D. L. Hartmann, 2001: The Madden–Julian oscillation, barotropic dynamics, and North Pacific tropical cyclone formation. Part I: Observations. J. Atmos. Sci., 58 , 25592570.

    • Search Google Scholar
    • Export Citation
  • Maloney, E. D., , and S. K. Esbensen, 2003: The amplification of east Pacific Madden–Julian oscillation convection and wind anomalies during June–November. J. Climate, 16 , 34823497.

    • Search Google Scholar
    • Export Citation
  • McGauley, M., , C. Zhang, , and N. Bond, 2004: Large-scale characteristics of the atmospheric boundary layer in the eastern Pacific cold tongue/ITCZ region. J. Climate, 17 , 39073920.

    • Search Google Scholar
    • Export Citation
  • Petersen, W. A., , R. Cifelli, , D. J. Boccippio, , ,S. A. Rutledge, , and C. Fairall, 2003: Convection and easterly wave structures observed in the eastern Pacific warm pool during EPIC-2001. J. Atmos. Sci., 60 , 17541773.

    • Search Google Scholar
    • Export Citation
  • Raymond, D. J., , G. B. Raga, , C. S. Bretherton, , J. Molinari, , C. López-Carillo, , and Ž Fuchs, 2003: Convective forcing in the intertropical convergence zone of the east Pacific. J. Atmos. Sci., 60 , 20642082.

    • Search Google Scholar
    • Export Citation
  • Raymond, D. J., and Coauthors, 2004: EPIC2001 and the coupled ocean–atmosphere system of the tropical east Pacific. Bull. Amer. Meteor. Soc., 85 , 13411354.

    • Search Google Scholar
    • Export Citation
  • Raymond, D. J., , C. S. Bretherton, , and J. Molinari, 2006: Dynamics of the intertropical convergence zone of the east Pacific. J. Atmos. Sci., in press.

    • Search Google Scholar
    • Export Citation
  • Serra, Y. L., , and R. A. Houze Jr., 2002: Observations of variability on synoptic time scales in the east Pacific ITCZ. J. Atmos. Sci., 59 , 17231743.

    • Search Google Scholar
    • Export Citation
  • Small, R. J., , S-P. Xie, , Y. Wang, , S. K. Esbensen, , and D. Vickers, 2005: Numerical simulation of boundary layer structure and cross-equatorial flow in the eastern Pacific. J. Atmos. Sci., 62 , 18121830.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., , and C-P. Chang, 1969: Spectrum analysis of large-scale wave disturbances in the tropical lower troposphere. J. Atmos. Sci., 26 , 10101025.

    • Search Google Scholar
    • Export Citation
  • Zhang, C., , M. McGauley, , and N. A. Bond, 2004: Shallow meridional circulation in the tropical eastern Pacific. J. Climate, 17 , 133139.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    The mean over all eight C-130 flights of (a) zonal wind, (b) meridional wind, and (c) pressure at the surface (solid line) and 1.6 km (dashed). ERA40 averages over the same eight days for the surface (circles) and 1.6 km (diamonds) are also shown. Winds on the eight flights were measured at approximately 30- and 1.6-km altitude. Pressures measured at flight level are hydrostatically adjusted to the surface and 1.6 km, and 1.6-km pressures are increased by 172.5 hPa for the plot.

  • View in gallery

    The first EOF (EOF1) of the perturbation pressure measured by the TAO buoys from May to Nov 2001 (thick dashed line), and EOF1 of the combined surface and 1.6-km perturbation pressure from the eight C-130 95°W flights (thin lines).

  • View in gallery

    Scatterplot of the C-130 vs the TAO leading principal component (PC1C-130 vs PC1) for the eight days of the 95°W flights. The axes are unitless standard deviations for each principal component. The top axis is PC1 restandardized for the Sep–Oct period.

  • View in gallery

    The time series of the TAO leading principal component PC1 (thick line) and GOES infrared brightness temperature TIR (thin line). The gray circles show PC1C-130 offset by the mean of PC1 for the eight flight days.

  • View in gallery

    (a) The Sep–Oct TAO and C-130 surface wind vectors regressed respectively on the TAO PC1 (standardized for Sep–Oct) and PC1C-130. (b) The C-130 wind vector at 1.6-km altitude regressed against PC1C-130. Note the difference in wind vector scales between (a) the surface regressions and (b) the 1.6-km regression. Circles on the southernmost vectors represent the standard error at all latitudes.

  • View in gallery

    Latitude–height regressions against PC1C-130 of the (a) specific humidity, (b) potential temperature, and (c) wind from dropwinsondes. The contour interval is indicated above each plot. Positive regressions are indicated by solid contours, and negative regressions by dashed contours. The zero contour is omitted. In (c) the zonal wind is indicated by contours (westerlies solid) and the meridional wind is indicated by arrows. Statistically significant regions (at 95% confidence) are shaded. Significant meridional wind regressions are indicated by bold arrows.

  • View in gallery

    Sep–Oct 2001 mean surface wind and sea level pressure from ERA40. GPCP precipitation is shaded.

  • View in gallery

    Surface wind and sea level pressure (lag) regressions with PC1 from ERA40 in Sep–Oct 2001. The regressions (a) lead PC1 by 1 day, (b) are synchronous with PC1, and (c) lag PC1 by 1 day. The contour interval for sea level pressure is 0.1 hPa, negative contours are dashed, and the zero contour is thick. The locations of the TAO buoys used to determine the PC1 time series in Fig. 2 are marked with triangles.

  • View in gallery

    Plot of 850-hPa wind and geopotential anomalies from ERA40 associated with a one standard deviation anomaly of PC1 in Sep–Oct 2001. The contour interval for the geopotential is 10 m2 s−2, negative contours are dashed, and the zero contour is thick.

  • View in gallery

    Divergence anomalies from ERA40 (contours) and precipitation anomalies from GPCP (shaded) associated with PC1 in Sep–Oct 2001. The contour interval for the divergence regression is 5 × 10−7 s−1. White contours are negative (convergence), and the zero contour is not shown.

  • View in gallery

    The summer SLP and surface wind regression with PC1 from ERA40. As in Fig. 8b but for 18 May–31 Oct 2001.

  • View in gallery

    The summer SLP and surface wind regression with PC1 from ERA40. As in Fig. 11 but only the variation associated with the trend is included.

  • View in gallery

    The detrended, 15-day low-pass filtered summer SLP and surface wind regression with PC1 from ERA40.

  • View in gallery

    The 15-day high-pass filtered summer SLP and surface wind regression with PC1 from ERA40.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 10 10 1
PDF Downloads 6 6 0

Variability in the Southerly Flow into the Eastern Pacific ITCZ

View More View Less
  • 1 International Pacific Research Center, University of Hawaii at Manoa, Honolulu, Hawaii
  • 2 Department of Atmospheric Sciences, University of Washington, Seattle, Washington
© Get Permissions
Full access

Abstract

During boreal summer and fall, there is a strong southerly boundary layer flow across the equator into the east Pacific intertropical convergence zone (ITCZ). The modulation of this flow on synoptic to seasonal time scales is studied using an index of meridional pressure difference between the equator and the ITCZ along 95°W. Two complementary datasets from the East Pacific Investigation of Climate (EPIC) are used to study eastern Pacific variability. Daily measurements of sea level pressure (SLP) from Tropical Atmosphere Ocean (TOA) array buoys from May to November 2001 provide temporal coverage, and eight flights by a C-130 aircraft during September to October 2001 document the associated modulation of lower tropospheric vertical structure.

The principal mode of variability of the perturbation SLP along 95°W from 1°S to 12°N, derived by principal component analysis from either the eight flights (PC1C-130) or from daily TAO buoy observations (PC1), explains 77% of the meridional pressure gradient variability. The pressure anomalies at 1.6 km are similar to those at the surface. The time series of the first mode of the TAO observations shows that most of the variance is in the 2–7-day range. Low pressure at 12°N is associated with southerly and westerly surface wind anomalies, and enhanced precipitation in the ITCZ. The depth of ITCZ convection is more strongly correlated to meridional wind above the planetary boundary layer (PBL) than to meridional wind within the PBL. There is little correlation of PBL meridional flow across the equator with ITCZ convection.

Regression of PC1C-130 against the 95°W cross sections observed by dropwinsondes released during the eight C-130 flights shows correlations of westerlies to positive PC1C-130 (low pressure at 12°N). Between the equator and 4°N, statistically significant northerlies just above the PBL at 1–2-km height and southerlies at 4 km are correlated with negative PC1C-130, having high SLP at 12°N, an anomalously weak meridional SLP gradient, and suppressed convection in the ITCZ.

PC1 is bandpass filtered and correlated with reanalysis fields to identify the structures that modulate meridional pressure gradients along 95°W. Most of the variability at periods less than 15 days is related to easterly waves. Seasonal trends in PC1 during May–October 2001 reflect the seasonal evolution of the sea and land surface temperatures. After the seasonal trend is removed, a geostrophic westerly jet at 12°N—probably related to the Madden–Julian oscillation—dominates PC1 variability on time scales longer than 15 days.

Corresponding author address: S. P. de Szoeke, International Pacific Research Center, School of Ocean and Earth Science and Technology, University of Hawaii at Manoa, 2525 Correa Road, Honolulu, HI 96822. Email: deszoeke@hawaii.edu

Abstract

During boreal summer and fall, there is a strong southerly boundary layer flow across the equator into the east Pacific intertropical convergence zone (ITCZ). The modulation of this flow on synoptic to seasonal time scales is studied using an index of meridional pressure difference between the equator and the ITCZ along 95°W. Two complementary datasets from the East Pacific Investigation of Climate (EPIC) are used to study eastern Pacific variability. Daily measurements of sea level pressure (SLP) from Tropical Atmosphere Ocean (TOA) array buoys from May to November 2001 provide temporal coverage, and eight flights by a C-130 aircraft during September to October 2001 document the associated modulation of lower tropospheric vertical structure.

The principal mode of variability of the perturbation SLP along 95°W from 1°S to 12°N, derived by principal component analysis from either the eight flights (PC1C-130) or from daily TAO buoy observations (PC1), explains 77% of the meridional pressure gradient variability. The pressure anomalies at 1.6 km are similar to those at the surface. The time series of the first mode of the TAO observations shows that most of the variance is in the 2–7-day range. Low pressure at 12°N is associated with southerly and westerly surface wind anomalies, and enhanced precipitation in the ITCZ. The depth of ITCZ convection is more strongly correlated to meridional wind above the planetary boundary layer (PBL) than to meridional wind within the PBL. There is little correlation of PBL meridional flow across the equator with ITCZ convection.

Regression of PC1C-130 against the 95°W cross sections observed by dropwinsondes released during the eight C-130 flights shows correlations of westerlies to positive PC1C-130 (low pressure at 12°N). Between the equator and 4°N, statistically significant northerlies just above the PBL at 1–2-km height and southerlies at 4 km are correlated with negative PC1C-130, having high SLP at 12°N, an anomalously weak meridional SLP gradient, and suppressed convection in the ITCZ.

PC1 is bandpass filtered and correlated with reanalysis fields to identify the structures that modulate meridional pressure gradients along 95°W. Most of the variability at periods less than 15 days is related to easterly waves. Seasonal trends in PC1 during May–October 2001 reflect the seasonal evolution of the sea and land surface temperatures. After the seasonal trend is removed, a geostrophic westerly jet at 12°N—probably related to the Madden–Julian oscillation—dominates PC1 variability on time scales longer than 15 days.

Corresponding author address: S. P. de Szoeke, International Pacific Research Center, School of Ocean and Earth Science and Technology, University of Hawaii at Manoa, 2525 Correa Road, Honolulu, HI 96822. Email: deszoeke@hawaii.edu

1. Introduction

In boreal summer and autumn, southeasterly winds bring cool air from the southeastern tropical Pacific across the equatorial cold tongue. As the winds flow northward over increasingly warmer water in the Northern Hemisphere, the planetary boundary layer (PBL) is warmed and moistened by surface heat flux and evaporation, and its capping inversion weakens. At 5°N cumulus congestus are common. The atmosphere undergoes sporadic deep convection in the ITCZ, a zonal belt of low-level convergence over the southern edge of the warmest sea surface temperature (SST) between 8° and 12°N. The upward mass flux and latent heat release in deep convective towers form the vigorous upward branch of the Hadley circulation, which transmits surface solar heating throughout the troposphere. The distribution of deep convection within the ITCZ is highly variable in time and space, and is determined by a number of factors. Atmospheric waves such as the westerly propagating Madden–Julian oscillation (MJO; Madden and Julian 1971) and easterly propagating tropical depressions (Wallace and Chang 1969; Serra and Houze 2002) modulate the convection in the eastern Pacific warm pool. One goal of the East Pacific Investigation of Climate (EPIC) 2001 field experiment (Raymond et al. 2004) was to better characterize the interplay of thermodynamics and dynamics in the southerly inflow from the equator into the ITCZ and how this inflow is modulated during the passage of ITCZ disturbances such as easterly waves.

Recent papers have documented the mean thermodynamic (de Szoeke et al. 2005) and dynamical (McGauley et al. 2004) structure of the inflow along 95°W using EPIC airborne and moored Tropical Atmosphere Ocean (TAO) array buoy observations. Further insight has been contributed by a range of models: comparison of EPIC observations with general circulation models (S. Esbensen 2003, personal communication), regional model simulations (Small et al. 2005), and large eddy simulations (de Szoeke and Bretherton 2004). Those papers provide context for this paper, which analyzes the atmospheric variability along 95°W. Hence we briefly review their findings.

Figure 1 shows the mean winds and pressure at 30 m and 1600 m above sea level from the EPIC airborne observations along 95°W (solid and dashed lines). Surface southeasterly winds turn to southwesterlies at 2°N, and there is a broad region of surface meridional convergence from 4° to 12°N, as shown in Figs. 1a and 1b. The mean surface winds represent an approximate balance among the surface and the entrainment drag, the Coriolis force, and the mean pressure gradient shown in Fig. 1c. The mean surface pressure gradient is about 3 hPa over 12° of latitude. At 1.6 km the mean pressure gradient and meridional wind are small, and there are easterlies of up to 5 m s−1. The strong surface pressure gradient and small pressure gradient above the PBL are congruent with the mechanism of Lindzen and Nigam (1987), which asserts that the pressure gradients at the surface are hydrostatically induced in the lower free troposphere by horizontal temperature gradients linked to the sea surface temperature.

Zhang et al. (2004) also noted in EPIC and other observations that there were slight mean northerlies at 2–4-km elevation south of the ITCZ and attributed this lower-tropospheric return flow to air diverging from the ITCZ that was not sufficiently warm and moist to rise in deep convection. Such a circulation could couple changes in the PBL winds and thermodynamic properties to the winds and humidity above the PBL.

The goal of this paper is to use EPIC observations to document and understand the atmospheric variability along the 95°W transect from the equator to the warm pool. This complements discussions by Raymond et al. (2003, 2006) of the variability of warm pool convection. Particular scientific issues include 1) What dynamical modes of variability most strongly affect the southerly inflow in this region? 2) Is enhanced ITCZ convection related to enhanced cross-equatorial flow? 3) Is surface pressure variability on short time scales consistent with the mechanism of Lindzen and Nigam (1987)? Since the pressure field is the link between the thermodynamic and dynamical structure of the inflow, we base our analysis on an index of the meridional surface pressure gradient. Variation of the vertical structure of the atmosphere was sampled by eight EPIC 2001 flights along 95°W. Although the EPIC 95°W flights are limited in number, the observations obtained on the flights are valuable for forming hypotheses, for testing models, and for planning any future field campaign in this remote location.

In section 2 we identify by principal component analysis the leading mode of pressure gradient variability in the EPIC 95°W flights and in the daily TAO buoy observations. The pressure variability is regressed against wind and against infrared brightness temperature—an indicator of ITCZ convection. In section 3 we show the vertical structure of variations in the circulation, temperature, and humidity along 95°W associated with the principal mode of pressure variability. In section 4 regressions with the European Centre for Medium-Range Weather Forecasts (ECMWF) 40-yr Reanalyses (ERA40) show the regional three-dimensional circulations associated with the observed variability. In section 5 this pressure variability is decomposed into seasonal, intraseasonal, and synoptic variations, and the circulations associated with each are analyzed. Section 6 summarizes the findings, and proposes some questions arising from them.

2. Principal component analysis of EPIC 95°W and TAO data

The TAO buoys along 95°W observed meteorological variables, including surface winds and atmospheric pressure, from their moorings at 2°S, 0°, 2°N, 3.5°N, 5°N, 8°N, 10°N, and 12°N (Cronin et al. 2002). The hourly TAO observations were averaged daily to remove the diurnal cycle. The TAO buoy data are available at the National Oceanic and Atmospheric Administration (NOAA) Pacific Marine Environmental Laboratory (PMEL) Web site, http://www.pmel.noaa.gov/tao/epic/.

We employ principal component analysis to investigate the meridional daily averaged sea level pressure (SLP) variability sampled by the TAO buoys along 95°W from May to November 2001. Because much of the TAO data from the equator and 2°N are missing due to vandalism of the TAO moorings, the first principal component was calculated from the 2°S, 3.5°N, 5°N, 8°N, 10°N, and 12°N buoys and then projected onto the available parts of the 0° and 2°N time series to get the full spatial pattern. For our analysis we use the perturbation pressure, calculated by first calculating the variation from the mean pressure at each location, then subtracting the meridional-mean pressure variation at each time, to focus on the meridionally varying pressure fluctuations.

The heavy dashed line of Fig. 2 shows the first EOF of the perturbation pressure from the TAO data. The EOF in Fig. 2 corresponds to the perturbation pressure associated with an anomaly of one standard deviation in the corresponding principal component. This mode is by far the most important mode of perturbation pressure variability, explaining 77% of the pressure perturbations in the TAO data. The pattern of the leading mode is essentially a meridional pressure gradient with low pressure to the north. One standard deviation of this leading mode results in an average pressure gradient of 8 Pa deg−1 between 2°S and 12°N, which would result in a northward meridional acceleration of 5.4 m s−1 day−1, if unbalanced by other forces. The pressure gradient associated with the first EOF is more concentrated to the north (8°–12°N), and weaker across the equator (2°S–4°N).

In September and October 2001, as a part of EPIC, the National Center for Atmospheric Research (NCAR) C-130 aircraft flew eight missions between 1°S and 12°N along the 95°W meridian (Raymond et al. 2004; McGauley et al. 2004; de Szoeke et al. 2005). The aircraft made horizontal legs at 30- and 1600-m altitude, and profiled the atmosphere on the intervening ascents and descents. The low 30-m aircraft legs were centered at 0.7°S, 0.7°N, 2.2°N, 3.7°N, 5.2°N, 6.7°N, 8.3°N, 10.3°N, and 12.2°N. The high (1.6 km) legs are leaved between the low legs at 0.1°S, 1.4°N, 3°N, 4.5°N, 5.9°N, 7.4°N, 9.2°N, and 11.3°N. Temperature, humidity, pressure, and velocities were sampled at a rate of 1 Hz by the aircraft. These data were downloaded from the University Corporation for Atmospheric Research (UCAR) Joint Office for Science Support (JOSS) Web site, http://www.joss.ucar.edu/epic/dm/data_access_frame.html.

As in Raymond et al. (2004), we averaged the pressure observations for each horizontal leg of each flight, hydrostatically adjusted the pressure data to the standard altitudes of 30 m and 1.6 km, and then removed the diurnal cycle from the pressure observations by subtracting the diurnal and semidiurnal harmonics of the pressure computed from hourly TAO surface pressure observations. We assume that the diurnal cycle is due to atmospheric tides, whose amplitudes are relatively uniform over the first 1.6-km height of the atmosphere and vary little between the equator and 12°N. Harmonics other than the diurnal and semidiurnal cycles are small. Figure 1 shows the meridional distribution of wind and pressure at 30 m and 1.6 km, averaged over all eight C-130 95°W flights. The perturbation pressure was computed for each location, as for the TAO data, before performing the principal component analysis.

The squares and triangles in Fig. 2 show the leading EOF of the perturbation pressure from the combined set of surface (30 m) and 1.6-km leg-averaged data for the C-130 95°W flights. The squares show the pressure EOF at the surface, and the triangles show the EOF at 1.6 km. The leading EOF of the C-130 data explains 77% of the C-130 pressure perturbations. The leading mode is associated with lower pressure to the north at both the surface and 1.6 km. The similarity of the pressure patterns at the two altitudes suggests that most of the variability in the pressure gradients is imposed from above the boundary layer, as also noted by Raymond et al. (2006). The surface pressure part of the EOF from the C-130 flattens off in the vicinity of the equator and has a gradient of about 1 × 10−3 hPa km−1 north of 4°N. At 1.6 km the gradient is somewhat more uniform with latitude, about 5 × 10−4 hPa km−1. The corresponding pressure difference of 0.65 hPa between the equator and 12°N corresponds to perturbation geostrophic westerlies of almost 5 m s−1.

The similarity between TAO and the C-130 EOFs in Fig. 2 means that it is possible, but not certain, that perturbation pressure variations observed by the two platforms result from the same phenomena. Figure 3 shows a strong correlation (r = 0.8) between the C-130 principal component time series (PC1C-130) and the May–November 2001 TAO principal component time series (PC1) at the C-130 sampling times, so it is likely that the variability in the C-130 observations is representative of the high-frequency variability in the TAO observations. Figure 4 shows that the EPIC 2001 intensive observation period, when the 95°W flights were made in late September and early October, was a period of particularly intense day-to-day variability in PC1 (thick line) and TIR (thin line). Based on the evidence of PC1C-130 (gray circles) in Fig. 4, we say this variability was sampled well by the eight C-130 flights.

Figure 4 shows the time series of the leading TAO principal component (PC1) and of the mean infrared brightness temperature TIR in the ITCZ study region (averaged over 8°–12°N, 93°–97°W from geostationary satellite measurements and obtained from D. Raymond) for the summer of 2001. Figure 4 shows that the infrared brightness temperature TIR is anticorrelated with PC1 (r = −0.54 for May–November); that is, lower pressure to the north coincides with active deep convection over the ITCZ study region.

The sudden shifts in the time series of PC1 and the Geostationary Operational Environmental Satellite (GOES) infrared brightness temperature (Fig. 4) at the end of May and the beginning of November show that a weak meridional pressure gradient and decreased ITCZ convection occur before and after the summer regime. We believe these shifts are associated with seasonal variation, though it is not clear why they should be so sudden.

From mid June to early November there is an increasing trend in PC1 that does not seem to be reflected in the infrared brightness temperature. In section 5 we will investigate the seasonal, intraseasonal, and daily variability exhibited by PC1 during June–October 2001.

Regressions of the wind from the C-130 flights along 95°W onto the C-130 first principal component (PC1C-130) are shown in Fig. 5. The regressions of the wind measured at the surface and 1.6 km are shown in Figs. 5a and 5b, respectively. South of 5°N, 1.6 km is in the free troposphere above the PBL capping inversion. North of 5°N a distinct capping inversion was not observed, and cumulus clouds of various sizes were encountered at this level. The wind signal is much stronger at 1.6 km than near the surface, presumably due to damping of boundary layer wind anomalies by the surface drag. Meridional convergence is comparably strong at the two levels, though with a different latitudinal structure. The low pressure to the north in the positive phase of the principal component is associated with an increase in surface southerlies of 1–2 m s−1 from 5° to 8°N and an increase in westerlies of 1 m s−1 between 3.5° and 10°N. There is a 1 m s−1 decrease in surface meridional wind between 8° and 10°N, corresponding to meridional convergence of 4 × 10−6 s−1. At 1.6 km, southerlies are enhanced by 1–4 m s−1 from the equator to 9°N. From 4° to 11°N the regressed westerly anomalies are 2–5 m s−1. The circle on the head of the arrows represents an estimated standard error of the regressed wind for all latitudes. The magnitude of the regressed winds are on the order of their standard error, indicating that the eight C-130 flights do not provide enough realizations to give statistical confidence in the wind regressions, though they do paint a picture dynamically consistent with the pressure variability.

In Fig. 5a we also used the 61 days of TAO wind data during September–October 2001 to regress the surface wind against PC1 with more confidence. Before performing the regressions, the mean of PC1 for September–October was removed, then PC1 was renormalized by dividing it by its standard deviation for the period (0.72). The C-130 and the TAO surface wind regressions of Fig. 5a agree qualitatively, but the C-130 wind vector magnitudes tend to be greater. PC1 is normalized for both platforms but, because of sampling differences between the C-130 and the TAO data, the standard deviations of the wind components are still different between the platforms. The eight C-130 flights sampled some strong events, even compared to the relatively strong variability in September and October (cf. Fig. 4), which could explain why the magnitudes of the C-130 wind regressions are twice as large as the TAO regressions from 5° to 8°N. Regressions of the TAO wind with PC1 are southwesterly at 3.5° and 5°N, southerly at 8°N, and westerly at 10°N. The correlation of the vector wind to PC1 is
i1520-0469-62-12-4400-e1
where the quantities denoted by σu, συ, and σPC1 are standard deviations. Note the numerator is the square root of the quadrature sum of the covariance of the wind components with PC1. The correlation of the wind vector to PC1 is 0.45 at 3.5°N, 0.49 at 5°N, 0.35 at 8°N, 0.41 at 10°N, and 0.13 at 12°N. We expect the correlation between the vector wind and PC1 to be at most 0.62, considering that PC1 explains only 0.77 of the meridional pressure gradient variance and that the wind responds also to the zonal pressure gradient, which is not measured. The correlation is further reduced because the wind is not necessarily in perfect balance with the pressure gradient. For 61 independent realizations, correlations greater than r = 0.25 are significant at the 95% level.

At 8° and 10°N, the changes in the mean winds associated with convection (TIR) or the meridional pressure gradient (PC1) are rather modest. Raymond et al. (2003), using a set of 20 EPIC aircraft missions—more than twice the number of missions used for Fig. 5a— found that convective rainfall in the eastern Pacific ITCZ is strongly correlated with wind speed (r = 0.66), presumably through the effect of wind speed on surface evaporation. Back and Bretherton (2005), using a much larger dataset—four years of daily satellite passive microwave retrievals—found a much weaker, though still statistically significant, correlation (r = 0.3) at 10°N, 95°W. Maloney and Esbensen (2003) found a similar correlation between convective activity and the local surface latent heat flux beneath it. Our results, based on two months of TAO data for September–October 2001, also suggest that there is a correlation of wind speed with convection associated with PC1. The highest correlation between PC1 and the daily averaged TAO surface wind speed is r = 0.53, occurring at 5°N. The strongest correlation between the TAO surface wind speed and the infrared brightness temperature TIR for September–October is r = −0.4 at 8°N, which is comparable to the correlation between PC1 and TIR (r = −0.33) during this period.

3. Vertical structures associated with the pressure variability

After sampling the troposphere from 30 m to 1.6 km, the C-130 flew back north along 95°W, releasing dropwinsondes from about 6-km altitude at 1° latitude intervals. The dropwinsonde observations provide altitude–latitude cross-sectional snapshots of the temperature, humidity, and wind along 95°W. The dropwinsonde profiles are vertically interpolated to a 40-m grid, then the mean and anomalies at each height and latitude are calculated. Regressing the anomalies from the eight dropwinsonde cross sections against the PC1C-130 time series yields the cross sections of specific humidity qυ, potential temperature θ, and wind components shown in Fig. 6.

Since there are only eight flights, the regressions in the cross sections are significant to 95% confidence in only a few regions, and even in these regions the significance could be merely fortuitous. (95% confidence requires a correlation coefficient of 0.7 in a two-sided Fisher test with 8 degrees of freedom.) Thus one must take a holistic view, examining the consistency of the fields with each other, to assess whether these results are meaningful. Though most of the pressure gradient associated with PC1C-130 is north of 5°N, the southerly meridional wind jet at 0°–4°N between 1 and 2 km is statistically significant. Both the regressed wind amplitude and its statistical significance are stronger above the PBL than in the PBL. The northerlies at 4 km, 0°–4°N were also significant, possibly indicating a time-varying return circulation correlated with PC1C-130 at this altitude. Westerlies are significant around 2 km at 9°–10°N, and are surrounded by a large region of westerlies of 4 m s−1 that are in approximate geostrophic balance with the meridional pressure anomalies associated with PC1C-130. The 2 m s−1 westerlies in the PBL capping inversion just below 1 km are also significant from 0° to 2°N.

The significant dipole with a 0.5-K warm anomaly above the PBL inversion, and a 0.5-K cold anomaly below the inversion around 3°N (Fig. 6b) indicates a combination of a stronger and a lower inversion for the high-index of PC1C-130. The PC1-regressed humidity (Fig. 6a) has a prominent (but not quite statistically significant) moist anomaly exceeding 1 g kg−1 in the anomalous northerly return flow. The moist anomaly does attain statistical significance at 2°N and 2.5-km altitude. This may be a signature of variations in ITCZ convective outflow, but one must note that air would take 4 days to traverse from 8°N to 2°N with a meridional speed of 2 m s−1, during which time PC1 would vary greatly. A careful trajectory analysis would be required to best interpret these correlations. There is a tendency for the boundary layer to be dry when PC1 is positive, possibly reflecting anomalous dry advection associated with slightly enhanced meridional wind speed. Though there are small patches of apparently significant correlations to PC1C-130 in the ITCZ region, these could be attributed to quickly evolving small-scale convective processes, which are only fortuitously correlated to PC1C-130.

We produced latitude–height regression plots like Fig. 6 from ERA40 limited to the eight flight days (not shown). The correspondence in zonal wind was quite good, but the potential temperature, meridional wind, and humidity anomalies bore little resemblance to the dropwinsonde observations, perhaps due to the weakness of the correlation of these fields to PC1. Whether due to sampling, ERA40 assimilation problems, or model error, this disagreement suggests a cautious approach to interpreting the details of the ERA40 PC1-regression maps that are presented in the next section.

4. Large-scale patterns from reanalysis

First we review the mean ERA40 wind and pressure patterns and compare them to the EPIC observations along 95°W. Figure 7 shows the mean ERA40 surface wind and SLP for September–October 2001. Surface winds are southerly, converging on the ITCZ. The pressure is lowest in or north of the ITCZ. Figure 7 also shows the September–October 2001 mean of the 1° × 1° daily Global Precipitation Climatology Project (GPCP) precipitation (shaded; Huffman et al. 2001), downloaded from the National Climatic Data Center Web site http://lwf.ncdc.noaa.gov/oa/wmo/wdcamet-ncdc.html. The ITCZ is demarcated by the band of precipitation and convergence at 8°–10°N. For comparison with the EPIC 95°W flight observations, Fig. 1 also includes the ERA40 10-m winds (circles, Figs. 1a and 1b) and 1.6-km winds (diamonds, Figs. 1a and 1b), and surface and 1.6-km pressure (Fig. 1c) averaged over the eight EPIC 95°W flight days. Most of the reanalysis fields agree with the observations. The most notable exception is the ERA40 zonal wind at 1.6 km, which is 1–3 m s−1 more westerly than the observations.

Figure 8b shows the September–October 2001 regression of ERA40 surface wind and SLP against the TAO PC1 renormalized for September–October 2001. The regressed SLP shows broad high pressure a few degrees south of the equator and a tightly confined globe of anomalous high pressure with a maximum of 0.4 hPa centered at 15°N, 108°W. The meridional pressure variation along 95°W from ERA40 is similar to the observed TAO EOF in Fig. 2, while zonal pressure gradients at 95°W are weak. The lowest pressure anomaly plotted (−0.8 hPa) is found over the Caribbean in a broad trough north of the mean ITCZ. A southwestward trough extends from the Caribbean to 10°N, 95°W. The strongest surface wind anomalies make a U-shaped jet in a mainly geostrophic response to the strong pressure gradients around this trough. South of the ITCZ the pressure anomaly shows substantial symmetry across the equator with small, but significant, westerly wind anomalies across the region. A weak zonal pressure gradient helps maintain the westerlies along the equator.

Figures 8a and 8c show regressions of ERA40 winds and SLP leading and lagging PC1 by one day, respectively. The SLP trough axis at 10°N is at 90°W for the regression leading PC1, at 95°W for the synchronous regression and at about 100°W for the regression lagging PC1. This corresponds to westward propagation at about 6 m s−1, consistent with the composite analysis of easterly waves observed at 10°N, 95°W by Petersen et al. (2003) from the NOAA ship Ronald H. Brown.

The pattern of 850-hPa geopotential (Fig. 9), representing the flow above the PBL, is qualitatively similar to the surface pressure pattern, indicating that the pressure anomalies associated with PC1 are barotropic through the lower troposphere in the reanalysis as well as in the C-130 EPIC 95°W observations. The winds at 850 hPa are stronger than those at the surface and more geostrophic over the ITCZ. Convergence in the ITCZ at 850 hPa (not shown) is less than half that at 1000–925 hPa (Fig. 10). East of 95°W and north of the equator, the 850-hPa winds head northeastward, joining the westerly ITCZ jet off the southwest coast of Nicaragua.

Most of the PBL convergence associated with PC1 occurs from 8°–10°N. Figure 10 shows the 1000–925 hPa average divergence (contours) from ERA40. A broad zonal band of convergence is located over the region of maximum surface pressure gradient at 8°–12°N. Along 95°W the rainfall anomaly (shaded in Fig. 10) is largest to the north of the maximum PBL convergence, centered over 13°N, 95°W. Here the PC1-regressed precipitation anomaly exceeds 5 mm day−1.

5. Modes of summer variability

The TAO pressure and ERA40 data extend over a longer time period than the September–October 2001 period that we have compared with the EPIC 95°W flights. Shifts in PC1 and the infrared brightness temperature TIR before 18 May and after 31 October (Fig. 4) respectively mark the beginning and end of a season whose variability appears statistically similar. We shall call this season from 18 May to 31 October the “summer” of 2001.

The regression of PC1 onto the ERA40 surface wind and SLP for the summer period is shown in Fig. 11. Although the trough at 10°N, 95°W is evident, the regression over the summer period is considerably more zonally symmetric than the regression for September–October (cf. Fig. 8b). The September–October regression in Fig. 8 sampled a period of intense westward-propagating waves associated with strong 4–10-day variability in PC1 and TIR. Other, lower frequency modes of variability contribute to the more zonal structure in Fig. 11, as we now discuss.

Several time scales of variability are evident in Fig. 4 during the summer period. First, there is an increasing trend in PC1 (two standard deviations over the 167 days), which does not seem to be reflected in TIR. Second, there is intraseasonal variability in both PC1 and TIR. Third, there are 4–10-day variations in PC1 and TIR associated with easterly waves, which often seem roughly periodic. Before performing the regressions, we isolate these three summer modes of variability by filtering the PC1 time series and the ERA40 SLP and 10-m winds into three components: 1) the 18 May–31 October trend, 2) the detrended and 15-day low-pass filtered PC1, and 3) the 15-day high-pass filtered PC1. These explain 33%, 34%, and 28% of the variance of PC1 over the summer period, respectively. The remaining 5% of the variance of PC1 over the summer is correlated between the 15-day low-pass and high-pass time series.

Figure 12 shows the spatial pattern corresponding to the summer trend in PC1. The lower pressure in the northern tropical Pacific and Caribbean where the SST is warming, and higher pressure in the southern and equatorial Pacific where SST is not warming can be regarded as a hydrostatic adjustment to the seasonal trend in the SST (not shown). The TAO buoy observations closely match the ERA40 trends.

Intraseasonal variation in PC1 in the summer of 2001 is characterized by anomalously convective periods at the end of May, the middle of July, and the end of September. We isolate intraseasonal variability by first detrending PC1 and then low-pass filtering with a 15-day period cutoff sixth-order Butterworth filter. The pattern of ERA40 intraseasonal SLP and wind variability regressed against the low-pass filtered PC1 (Fig. 13) is the most zonally symmetric of the patterns associated with the three modes of variability. This mode is associated with enhanced meridional SLP gradient at 12°–15°N, geostrophic westerlies in the ITCZ region, and small Ekman convergence on the north side of the ITCZ in ERA40. The enhanced westerlies are consistent with the intraseasonal variability described by Maloney and Hartmann (2001).

The PC1 variability during September–October was predominantly on periods shorter than 15 days. Correspondingly, the trough of low surface pressure centered over 10°–15°N, 95°W seen in the September–October regression with PC1 in Fig. 8b is evident in the 15-day high-pass filtered regression of PC1 on SLP over the whole summer shown in Fig. 14. PC1 also projects onto high pressure at 105°W as in Fig. 8b. However, the size of this anomaly in the summer regression is 1/4 its size in the September–October regression, and variability in the September–October period is responsible for almost all the contribution to the summer dipole of high SLP at 95°W and low SLP at 105°W. Thus this 105°W anomaly is not a robust feature of the entire summer and may be associated with a few nascent tropical cyclones.

6. Conclusions

Dynamical variability in the tropical east Pacific atmosphere is explored using observations along 95°W from May to October 2001. The leading principal component (PC1) of the daily pressure perturbation along 95°W between the equator and 12°N explains 77% of the meridional pressure gradient variability there. The spatial pattern of the pressure variability from the EPIC 95°W flights is the same at the surface and at 1.6 km, having low pressure at 12°N and high pressure at the equator, spanned by a meridional pressure gradient of −7 × 10−4 hPa km−1, for one standard deviation of PC1. An enhanced meridional pressure gradient along 95°W is correlated to higher surface wind speed and more rainfall in and north of the ITCZ.

Winds from dropwinsondes regressed on the principal component PC1C-130 of the surface pressure from the C-130 show westerlies over most of 95°W between the equator and 12°N from the surface to 5-km altitude. Westerly anomalies are particularly strong in the ITCZ, consistent with the mechanism of Maloney and Hartmann (2001), in which the zonal wind anomaly contributes to the eddy kinetic energy, feeding synoptic disturbances. The winds in the PBL respond less strongly to variations in PC1C-130 than the winds in the free troposphere, presumably due to frictional damping in the PBL. Meridional flow in the PBL south of 5°N is not well correlated to PC1 or ITCZ convection.

Our analysis also suggests that the shallow return circulation above the cross-equatorial southerly PBL, found in the mean along 95°W by Zhang et al. (2004), is modulated by PC1. The northerly return flow is elevated around 4 km in convectively active periods (positive PC1), with positive moisture anomalies above 2 km. In suppressed (negative PC1) periods the northerly return flow is just above the boundary layer capping inversion, at 1–2 km. Near the equator the inversion is slightly lower and sharper for positive PC1C-130 in the dropwinsonde observations. The observations during EPIC were only barely sufficient to determine the vertical structure of the potential temperature and humidity perturbations associated with PC1C-130 with statistical significance, so conclusions drawn from these regressions may not all be robust.

Zhang et al. (2004) proposes that the shallow circulation can affect the PBL wind through entrainment of momentum at the PBL top. We propose that in addition, variations in the meridional pressure gradient can change the wind in the PBL. We find variations of the flow above the boundary layer to be correlated with pressure gradient variations. Nevertheless, the weak correlation between the PBL wind and PC1 seen here supports neither the entrainment mechanism nor the pressure gradient mechanism directly.

Based on the time series of PC1, we decompose the observed pressure variability over the summer of 2001 into three types: 1) a trend due to a drop in surface pressure during the summer where the surface temperature is warming, 2) the intraseasonal variation in the geostrophic westerlies in the ITCZ latitudes, and 3) the weekly and subweekly variations associated with easterly waves in the ITCZ. The easterly waves dominated the PC1 variability during EPIC 2001 (September–October 2001), but intraseasonal variability and the trend are also important, especially over the entire summer.

Sudden shifts in the PC1 and the infrared brightness temperature in mid May and early November mark the ends of the convectively active summer for the northern ITCZ. At 10°N the shifts are most pronounced in the far eastern Pacific. Similar shifts at almost the same time of year are seen in the outgoing longwave radiation (OLR) at 10°N for 2000 and 2002. Whether such shifts are seen in other years, why they are so sudden, and how they are related the march of the seasonal cycle are questions for further study.

Except for the trend in PC1, surface pressure variability is not governed by a hydrostatic response to surface temperature. While the hydrostatic response of the surface pressure to surface temperature (Lindzen and Nigam 1987) is important in setting the seasonal-mean pressure gradients, as shown by Raymond et al. (2006), tropospheric dynamics above the PBL are responsible for intraseasonal and synoptic surface pressure variability.

The variation of the zonal wind tends toward geostrophy with the meridional pressure gradient, but the covariation of the other fields is more subtle. For an enhanced pressure gradient, significant jets, northward toward the ITCZ at 2-km and southward at 4-km altitude, are observed along with a dry PBL and a moist free troposphere. This meridional circulation is likely connected to convective entrainment and detrainment in the ITCZ; its relation to ITCZ convection and to equatorial waves should be further explored. The return circulation aloft helps moisten air entrained into the PBL as far as 5°S, as seen on the EPIC 2001 stratocumulus cruise (Bretherton et al. 2004), and might interact with boundary layer cloudiness in this region (de Szoeke and Bretherton 2004). ERA40 trajectories, and rawinsonde and profiler data from the Galapagos Islands (Hartten and Gage 2000) may provide useful long-term records for analysis of the return circulation.

Acknowledgments

Spirited discussion among the authors, David Raymond, and John Molinari inspired much of this investigation. We thank David Raymond and Beverly Davis for providing the GOES infrared brightness temperature data, the NCAR Research Aviation Facility and the TAO office for providing observational data for the EPIC project, and the NCAR Data Support Section and ECMWF for providing the reanalysis data. We acknowledge support from NSF Grants ATM-0082384, ATM-0082391, and ATM-0433712.

REFERENCES

  • Back, L. E., , and C. S. Bretherton, 2005: The relationship between wind speed and precipitation in the Pacific ITCZ. J. Climate, 18 , 43174328.

    • Search Google Scholar
    • Export Citation
  • Bretherton, C. S., , T. Uttal, , C. W. Fairall, , S. Yuter, , R. Weller, , D. Baumgardner, , K. Comstock, , and R. Wood, 2004: The EPIC 2001 stratocumulus study. Bull. Amer. Meteor. Soc., 85 , 967977.

    • Search Google Scholar
    • Export Citation
  • Cronin, M. F., , N. A. Bond, , C. Fairall, , J. Hare, , M. J. McPhaden, , and R. A. Weller, 2002: Enhanced oceanic and atmospheric monitoring underway in eastern Pacific. Eos, Trans. Amer. Geophys. Union, 83 , 210211.

    • Search Google Scholar
    • Export Citation
  • de Szoeke, S. P., , and C. S. Bretherton, 2004: Quasi-Lagrangian large eddy simulations of cross-equatorial flow in the east Pacific atmospheric boundary layer. J. Atmos. Sci., 61 , 18371858.

    • Search Google Scholar
    • Export Citation
  • de Szoeke, S. P., , C. S. Bretherton, , N. A. Bond, , M. F. Cronin, , and B. M. Morley, 2005: EPIC 95°W observations of the eastern Pacific atmospheric boundary layer from the cold tongue to the ITCZ. J. Atmos. Sci., 62 , 426442.

    • Search Google Scholar
    • Export Citation
  • Hartten, L. M., , and K. S. Gage, 2000: ENSO’s impact on the annual cycle: The view from Galapagos. Geophys. Res. Lett., 27 , 385388.

  • Huffman, G. J., and Coauthors, 2001: Global precipitation at one-degree daily resolution from multisatellite observations. J. Hydrometeor., 2 , 3650.

    • Search Google Scholar
    • Export Citation
  • Lindzen, R. S., , and S. Nigam, 1987: On the role of sea surface temperature gradients in forcing low-level winds and convergence in the tropics. J. Atmos. Sci., 44 , 24182436.

    • Search Google Scholar
    • Export Citation
  • Madden, R. A., , and P. R. Julian, 1971: Detection of a 40–50 day oscillation in the zonal wind in the tropical Pacific. J. Atmos. Sci., 28 , 702708.

    • Search Google Scholar
    • Export Citation
  • Maloney, E. D., , and D. L. Hartmann, 2001: The Madden–Julian oscillation, barotropic dynamics, and North Pacific tropical cyclone formation. Part I: Observations. J. Atmos. Sci., 58 , 25592570.

    • Search Google Scholar
    • Export Citation
  • Maloney, E. D., , and S. K. Esbensen, 2003: The amplification of east Pacific Madden–Julian oscillation convection and wind anomalies during June–November. J. Climate, 16 , 34823497.

    • Search Google Scholar
    • Export Citation
  • McGauley, M., , C. Zhang, , and N. Bond, 2004: Large-scale characteristics of the atmospheric boundary layer in the eastern Pacific cold tongue/ITCZ region. J. Climate, 17 , 39073920.

    • Search Google Scholar
    • Export Citation
  • Petersen, W. A., , R. Cifelli, , D. J. Boccippio, , ,S. A. Rutledge, , and C. Fairall, 2003: Convection and easterly wave structures observed in the eastern Pacific warm pool during EPIC-2001. J. Atmos. Sci., 60 , 17541773.

    • Search Google Scholar
    • Export Citation
  • Raymond, D. J., , G. B. Raga, , C. S. Bretherton, , J. Molinari, , C. López-Carillo, , and Ž Fuchs, 2003: Convective forcing in the intertropical convergence zone of the east Pacific. J. Atmos. Sci., 60 , 20642082.

    • Search Google Scholar
    • Export Citation
  • Raymond, D. J., and Coauthors, 2004: EPIC2001 and the coupled ocean–atmosphere system of the tropical east Pacific. Bull. Amer. Meteor. Soc., 85 , 13411354.

    • Search Google Scholar
    • Export Citation
  • Raymond, D. J., , C. S. Bretherton, , and J. Molinari, 2006: Dynamics of the intertropical convergence zone of the east Pacific. J. Atmos. Sci., in press.

    • Search Google Scholar
    • Export Citation
  • Serra, Y. L., , and R. A. Houze Jr., 2002: Observations of variability on synoptic time scales in the east Pacific ITCZ. J. Atmos. Sci., 59 , 17231743.

    • Search Google Scholar
    • Export Citation
  • Small, R. J., , S-P. Xie, , Y. Wang, , S. K. Esbensen, , and D. Vickers, 2005: Numerical simulation of boundary layer structure and cross-equatorial flow in the eastern Pacific. J. Atmos. Sci., 62 , 18121830.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., , and C-P. Chang, 1969: Spectrum analysis of large-scale wave disturbances in the tropical lower troposphere. J. Atmos. Sci., 26 , 10101025.

    • Search Google Scholar
    • Export Citation
  • Zhang, C., , M. McGauley, , and N. A. Bond, 2004: Shallow meridional circulation in the tropical eastern Pacific. J. Climate, 17 , 133139.

    • Search Google Scholar
    • Export Citation
Fig. 1.
Fig. 1.

The mean over all eight C-130 flights of (a) zonal wind, (b) meridional wind, and (c) pressure at the surface (solid line) and 1.6 km (dashed). ERA40 averages over the same eight days for the surface (circles) and 1.6 km (diamonds) are also shown. Winds on the eight flights were measured at approximately 30- and 1.6-km altitude. Pressures measured at flight level are hydrostatically adjusted to the surface and 1.6 km, and 1.6-km pressures are increased by 172.5 hPa for the plot.

Citation: Journal of the Atmospheric Sciences 62, 12; 10.1175/JAS3626.1

Fig. 2.
Fig. 2.

The first EOF (EOF1) of the perturbation pressure measured by the TAO buoys from May to Nov 2001 (thick dashed line), and EOF1 of the combined surface and 1.6-km perturbation pressure from the eight C-130 95°W flights (thin lines).

Citation: Journal of the Atmospheric Sciences 62, 12; 10.1175/JAS3626.1

Fig. 3.
Fig. 3.

Scatterplot of the C-130 vs the TAO leading principal component (PC1C-130 vs PC1) for the eight days of the 95°W flights. The axes are unitless standard deviations for each principal component. The top axis is PC1 restandardized for the Sep–Oct period.

Citation: Journal of the Atmospheric Sciences 62, 12; 10.1175/JAS3626.1

Fig. 4.
Fig. 4.

The time series of the TAO leading principal component PC1 (thick line) and GOES infrared brightness temperature TIR (thin line). The gray circles show PC1C-130 offset by the mean of PC1 for the eight flight days.

Citation: Journal of the Atmospheric Sciences 62, 12; 10.1175/JAS3626.1

Fig. 5.
Fig. 5.

(a) The Sep–Oct TAO and C-130 surface wind vectors regressed respectively on the TAO PC1 (standardized for Sep–Oct) and PC1C-130. (b) The C-130 wind vector at 1.6-km altitude regressed against PC1C-130. Note the difference in wind vector scales between (a) the surface regressions and (b) the 1.6-km regression. Circles on the southernmost vectors represent the standard error at all latitudes.

Citation: Journal of the Atmospheric Sciences 62, 12; 10.1175/JAS3626.1

Fig. 6.
Fig. 6.

Latitude–height regressions against PC1C-130 of the (a) specific humidity, (b) potential temperature, and (c) wind from dropwinsondes. The contour interval is indicated above each plot. Positive regressions are indicated by solid contours, and negative regressions by dashed contours. The zero contour is omitted. In (c) the zonal wind is indicated by contours (westerlies solid) and the meridional wind is indicated by arrows. Statistically significant regions (at 95% confidence) are shaded. Significant meridional wind regressions are indicated by bold arrows.

Citation: Journal of the Atmospheric Sciences 62, 12; 10.1175/JAS3626.1

Fig. 7.
Fig. 7.

Sep–Oct 2001 mean surface wind and sea level pressure from ERA40. GPCP precipitation is shaded.

Citation: Journal of the Atmospheric Sciences 62, 12; 10.1175/JAS3626.1

Fig. 8.
Fig. 8.

Surface wind and sea level pressure (lag) regressions with PC1 from ERA40 in Sep–Oct 2001. The regressions (a) lead PC1 by 1 day, (b) are synchronous with PC1, and (c) lag PC1 by 1 day. The contour interval for sea level pressure is 0.1 hPa, negative contours are dashed, and the zero contour is thick. The locations of the TAO buoys used to determine the PC1 time series in Fig. 2 are marked with triangles.

Citation: Journal of the Atmospheric Sciences 62, 12; 10.1175/JAS3626.1

Fig. 9.
Fig. 9.

Plot of 850-hPa wind and geopotential anomalies from ERA40 associated with a one standard deviation anomaly of PC1 in Sep–Oct 2001. The contour interval for the geopotential is 10 m2 s−2, negative contours are dashed, and the zero contour is thick.

Citation: Journal of the Atmospheric Sciences 62, 12; 10.1175/JAS3626.1

Fig. 10.
Fig. 10.

Divergence anomalies from ERA40 (contours) and precipitation anomalies from GPCP (shaded) associated with PC1 in Sep–Oct 2001. The contour interval for the divergence regression is 5 × 10−7 s−1. White contours are negative (convergence), and the zero contour is not shown.

Citation: Journal of the Atmospheric Sciences 62, 12; 10.1175/JAS3626.1

Fig. 11.
Fig. 11.

The summer SLP and surface wind regression with PC1 from ERA40. As in Fig. 8b but for 18 May–31 Oct 2001.

Citation: Journal of the Atmospheric Sciences 62, 12; 10.1175/JAS3626.1

Fig. 12.
Fig. 12.

The summer SLP and surface wind regression with PC1 from ERA40. As in Fig. 11 but only the variation associated with the trend is included.

Citation: Journal of the Atmospheric Sciences 62, 12; 10.1175/JAS3626.1

Fig. 13.
Fig. 13.

The detrended, 15-day low-pass filtered summer SLP and surface wind regression with PC1 from ERA40.

Citation: Journal of the Atmospheric Sciences 62, 12; 10.1175/JAS3626.1

Fig. 14.
Fig. 14.

The 15-day high-pass filtered summer SLP and surface wind regression with PC1 from ERA40.

Citation: Journal of the Atmospheric Sciences 62, 12; 10.1175/JAS3626.1

* School of Ocean and Earth Science and Technology Contribution Number 6629 and International Pacific Research Center Contribution Number IPRC-342.

Save