• Adler, R. F., and Coauthors, 2003: The version-2 Global Precipitation Climatology Project (GPCP) monthly precipitation analysis (1979–present). J. Hydrometeor., 4, 11471167.

    • Search Google Scholar
    • Export Citation
  • Awaka, J., , T. Iguchi, , and K. Okamoto, 1998: Early results on rain type classification by the Tropical Rainfall Measuring Mission (TRMM) precipitation radar. Proc. Eighth URSI Commission F Open Symp., Averior, Portugal, URSI, 134–146.

    • Search Google Scholar
    • Export Citation
  • Back, L. E., , and C. S. Bretherton, 2006: Geographic variability in the export of moist static energy and vertical motion profiles in the tropical Pacific. Geophys. Res. Lett., 33, L17810, doi:10.1029/2006GL026672.

    • Search Google Scholar
    • Export Citation
  • Berg, W., , C. Kummerow, , and C. Morales, 2002: Differences between east and west Pacific rainfall systems. J. Climate, 15, 36593672.

  • Cess, R. D., , M. Zhang, , B. A. Wielicki, , D. F. Young, , X. Zhou, , and Y. Nikitenko, 2001: The influence of the 1998 El Niño upon cloud radiative forcing over the Pacific warm pool. J. Climate, 14, 21292137.

    • Search Google Scholar
    • Export Citation
  • Chen, Y., , A. D. Del Genio, , and J. Chen, 2007: The tropical atmospheric El Niño signal in satellite precipitation data and a global climate model. J. Climate, 20, 35803601.

    • Search Google Scholar
    • Export Citation
  • Curtis, S., , and R. Adler, 2000: ENSO indices based on patterns of satellite-derived precipitation. J. Climate, 13, 27862793.

  • Dai, A., , and T. M. L. Wigley, 2000: Global patterns of ENSO-induced precipitation. Geophys. Res. Lett., 27, 12831286.

  • Del Genio, A. D., , and W. Kovari, 2002: Climatic properties of tropical precipitating convection under varying environmental conditions. J. Climate, 15, 25972615.

    • Search Google Scholar
    • Export Citation
  • DeMoss, J. D., , and K. P. Bowman, 2007: Changes in TRMM Rainfall due to the orbit boost estimated from buoy rain gauge data. J. Atmos. Oceanic Technol., 24, 15981607.

    • Search Google Scholar
    • Export Citation
  • Fu, Y., , and G. Liu, 2001: The variability of tropical precipitation profiles and its impact on microwave brightness temperatures as inferred from TRMM data. J. Appl. Meteor., 40, 21302143.

    • Search Google Scholar
    • Export Citation
  • Graham, N. E., 1995: Simulation of recent global temperature trends. Science, 267, 666671.

  • Houze, R. A., Jr., 1993: Cloud Dynamics. Academic Press, 573 pp.

  • Houze, R. A., Jr., 1997: Stratiform precipitation in regions of convection: A meteorological paradox? Bull. Amer. Meteor. Soc., 78, 21792196.

    • Search Google Scholar
    • Export Citation
  • Iguchi, T., , T. Kozu, , R. Meneghini, , J. Awaka, , and K. Okamoto, 2000: Rain-profiling algorithm for the TRMM Precipitation Radar. J. Appl. Meteor., 39, 20382052.

    • Search Google Scholar
    • Export Citation
  • Kubar, T. L., , and D. L. Hartmann, 2008: The vertical structure of tropical oceanic convective clouds and its relation to precipitation. Geophys. Res. Lett., 35, L03804, doi:10.1029/2007GL032811.

    • Search Google Scholar
    • Export Citation
  • Lambert, F. H., , S. Alexander, , Y. K. Nir, , and J. Chiang, 2008: How much will precipitation increase with global warming? Eos, Trans. Amer. Geophys. Union, 89, 21, doi:10.1029/2008EO210001.

    • Search Google Scholar
    • Export Citation
  • Li, R., , and Q.-L. Min, 2010: The impacts of mineral dust on the vertical structure of precipitation. J. Geophys. Res., 115, D09203, doi:10.1029/2009JD011925.

    • Search Google Scholar
    • Export Citation
  • Li, R., , Y. Fu, , and P. Zhao, 2005: Characteristics of rainfall structure over the tropical Pacific during the later period of 1997/1998 El Niño derived from TRMM PR observations (in Chinese). Chin. J. Atmos. Sci., 29, 225235.

    • Search Google Scholar
    • Export Citation
  • Lin, B., , B. A. Wielicki, , P. Minnis, , L. Chambers, , K. M. Xu, , Y. Hu, , and A. Fan, 2006: The effect of environmental conditions on tropical deep convective systems observed from the TRMM satellite. J. Climate, 19, 57455761.

    • Search Google Scholar
    • Export Citation
  • Liu, G., , and Y. Fu, 2001: The characteristics of tropical precipitation profiles as inferred from satellite radar measurements. J. Meteor. Soc. Japan, 79, 131143.

    • Search Google Scholar
    • Export Citation
  • Mapes, B. E., , and R. A. Houze Jr., 1995: Diabatic divergence profiles in western Pacific mesoscale convective systems. J. Atmos. Sci., 52, 18071828.

    • Search Google Scholar
    • Export Citation
  • Min, Q.-L., , R. Li, , B. Lin, , E. Joseph, , S. Wang, , Y. Hu, , V. Morris, , and F. Chang, 2009: Evidence of mineral dust altering cloud microphysics and precipitation. Atmos. Chem. Phys., 9, 32233231.

    • Search Google Scholar
    • Export Citation
  • Reynolds, R. W., , N. A. Rayner, , T. M. Smith, , D. C. Stokes, , and W. Wang, 2002: An improved in situ and satellite SST analysis for climate. J. Climate, 15, 16091625.

    • Search Google Scholar
    • Export Citation
  • Rogers, R. R., , and M. K. Yau, 1989: A Short Course in Cloud Physics. 3rd ed. Pergamon Press, 293 pp.

  • Ropelewski, C., , and M. Halpert, 1987: Global and regional scale precipitation patterns associated with the El Niño/Southern Oscillation. Mon. Wea. Rev., 115, 16061626.

    • Search Google Scholar
    • Export Citation
  • Satoh, S., , and A. Noda, 2001: Retrieval of latent heating profiles from TRMM radar data. Preprints, 30th Int. Conf. on Radar Meteorology, Munich, Germany, Amer. Meteor. Soc., 6.3. [Available online at http://ams.confex.com/ams/30radar/techprogram/paper_21763.htm.]

    • Search Google Scholar
    • Export Citation
  • Schumacher, C., , and R. A. Houze Jr., 2003: Stratiform rain in the tropics as seen by the TRMM precipitation radar. J. Climate, 16, 17391756.

    • Search Google Scholar
    • Export Citation
  • Schumacher, C., , R. A. Houze Jr., , and I. Kraucunas, 2004: The tropical dynamical response to latent heating estimates derived from the TRMM precipitation radar. J. Atmos. Sci., 61, 13411358.

    • Search Google Scholar
    • Export Citation
  • Shin, D.-B., , G. R. North, , and K. P. Bowman, 2000: A summary of reflectivity profiles from the first year of TRMM radar data. J. Climate, 13, 40724086.

    • Search Google Scholar
    • Export Citation
  • Short, D. A., , and K. Nakamura, 2000: TRMM radar observations of shallow precipitation over the tropical oceans. J. Climate, 13, 41074124.

    • Search Google Scholar
    • Export Citation
  • Soden, B. J., 2000: The sensitivity of the tropical hydrological cycle to ENSO. J. Climate, 13, 538549.

  • Solomon, S., , D. Qin, , M. Manning, , M. Marquis, , K. Averyt, , M. M. B. Tignor, , H. L. Miller Jr., , and Z. Chen, Eds., 2007: Climate Change 2007: The Physical Science Basis. Cambridge University Press, 996 pp.

    • Search Google Scholar
    • Export Citation
  • Steiner, M., , R. A. Houze Jr., , and S. E. Yuter, 1995: Climatological characterization of three-dimensional storm structure from operational radar and rain gauge data. J. Appl. Meteor., 34, 19782007.

    • Search Google Scholar
    • Export Citation
  • Tao, W. K., , J. Simpson, , S. Lang, , M. McCumber, , R. Adler, , and R. Penc, 1990: An algorithm to estimate the heating budget from vertical hydrometeor profiles. J. Appl. Meteor., 29, 12321244.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., 1998: Atmospheric moisture residence times and cycling: Implications for rainfall rates with climate change. Climatic Change, 39, 667694.

    • Search Google Scholar
    • Export Citation
  • Wang, P. H., , P. Minnis, , B. A. Wielicki, , T. Wong, , R. D. Cess, , M. H. Zhang, , L. B. Vann, , and G. S. Kent, 2003: Characteristics of the 1997/1998 El Niño cloud distributions from SAGE II observations. J. Geophys. Res., 108, 4009, doi:10.1029/2002JD002501.

    • Search Google Scholar
    • Export Citation
  • Wentz, F. J., , L. Ricciardulli, , K. Hilburn, , and C. Mears, 2007: How much more rain will global warming bring? Science, 317, 233235, doi:10.1126/science.1140746.

    • Search Google Scholar
    • Export Citation
  • Yang, S., , and E. A. Smith, 1999: Moisture budget analysis of TOGA COARE area using SSM/I-retrieved latent heating and large-scale Q2 estimates. J. Atmos. Oceanic Technol., 16, 633655.

    • Search Google Scholar
    • Export Citation
  • Yulaeva, E., , and J. M. Wallace, 1994: The signature of ENSO in global temperature and precipitation fields derived from the Microwave Sounding Unit. J. Climate, 7, 17191736.

    • Search Google Scholar
    • Export Citation
  • Zhang, C., 2001: Double ITCZs. J. Geophys. Res., 106 (D11), 11 78511 792.

  • View in gallery

    Mean and anomalies of (a) sea surface temperature, (b) vertical velocity (Omega) at 500 hPa, and (c) RH at 500 hPa in tropical Pacific in 1998JFMA. Two red boxes represent the selected east and west Pacific areas. Accompanied with associated area mean vertical profiles in 1998JFMA (red curves) and 9900JFMA (blue curves).

  • View in gallery

    Comparison of monthly mean convective rain properties in tropical Pacific in 1998JFMA with those in 9900JFMA: (a),(d) the near-surface rain rate; (b),(e) the storm height; and (c),(f) the ratio of difference between rain rate at 6 and 4 km [CR(4km)-CR(6km)] to rain rate at 2 km [CR(2km)], where CR stands for convective rains.

  • View in gallery

    As in Fig. 2, but for stratiform rain.

  • View in gallery

    PDF of storm height in the selected (a) east and (b) west Pacific areas in 1998JFMA and 9900JFMA. All types of rain detected by TRMM PR are included.

  • View in gallery

    Mean relationship between precipitation-top height and the near-surface rain rate in EP observed by TRMM PR.

  • View in gallery

    Scatterplots of SST to PTH for (a) convective rain and (b) stratiform rain. Dark gray cross for 1998 JFMA; light gray cross for 9900JFMA. The associated mean relationships (0.5°C SST interval) with standard deviation for 1998JFMA and 9900JFMA are overlapped. Additionally, the associated linear regression results are also overlapped: the gray lines stand for 1998JFMA, dark gray lines for 9900JFMA, and black thick lines stand for the whole 12 months.

  • View in gallery

    Mean PTT under given SST and given (a),(d) Omega, (b),(e) RH, and (c),(f) the divergence at 200 mb for (a)–(c) convective and (d)–(f) stratiform rains in 1998–2000 JFMA in EP.

  • View in gallery

    (a) Overall mean rain rate profiles, (b) mean profiles for convective rains, and (c) for stratiform rains under given PTT and Rsrf in EP.

  • View in gallery

    (a) The mean SlopeA, SlopeB, and SlopeC as functions of surface rain rate and precipitation-top temperature in (left) 1998JFMA and in (right) 9900JFMA for convective rains. (b) As in (a), but for stratiform rains.

  • View in gallery

    The mean rain rate profiles of (a) convective, (b) stratiform, and (c) warm rains normalized by Rsrf and (d)–(f) the associated vertical gradient profiles (times by 250 m).

  • View in gallery

    (a) The estimated profiles of LH index in 1998JFMA in EP based on three tests (T1, T2, T3). (b) The relative difference among those three results.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 23 23 6
PDF Downloads 8 8 3

1997/98 El Niño–Induced Changes in Rainfall Vertical Structure in the East Pacific

View More View Less
  • 1 Atmospheric Sciences Research Center, State University of New York at Albany, Albany, New York
  • 2 Laboratory of Satellite Remote Sensing and Climate Environment, University of Science and Technology of China, and State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics (LASG), Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
© Get Permissions
Full access

Abstract

The 1997/98 El Niño–induced changes in rainfall vertical structure in the east Pacific (EP) are investigated by using collocated Tropical Rainfall Measuring Mission (TRMM) precipitation radar (PR) and associated daily SST and 6-hourly reanalysis data during January, February, March, and April of 1998, 1999, and 2000. This study shows that there are five key parameters, that is, surface rain rate, precipitation-top height (or temperature), and precipitation growth rates at upper, middle, and low layers to define a rainfall profile, and those five key parameters are strongly influenced by both SST and large-scale dynamics. Under the influence of 1997/98 El Niño, the precipitation-top heights in the EP were systematically higher by about 1 km than those under non–El Niño conditions, while the freezing level was about 0.5 km higher. Under the constraints of rain type, surface rain rate, and the precipitation top, the shape of rainfall profile still showed significant differences: the rain growth was relatively faster in the mid-layer (−5° to +2°C isotherm) but slower in the lower layer (below +2°C isotherm) under the influence of El Niño. It is also evident that the dependence of precipitation top height on SST was stronger under large-scale decent (non–El Niño) circulations but much weaker under large-scale ascent (El Niño) circulations. The combined effect of larger vertical extent and greater growth rate in the middle layer further shifted latent heating upward as compared with the impact of horizontal changes in the rain type fractions (convective versus stratiform). Such additional latent heating shift would certainly further elevate circulation centers and strengthen the upper-layer circulation.

Corresponding author address: Qilong Min, Atmospheric Sciences Research Center, State University of New York at Albany, 1400 Washington Avenue, Albany, NY 12222. E-mail: min@asrc.cestm.albany.edu

Abstract

The 1997/98 El Niño–induced changes in rainfall vertical structure in the east Pacific (EP) are investigated by using collocated Tropical Rainfall Measuring Mission (TRMM) precipitation radar (PR) and associated daily SST and 6-hourly reanalysis data during January, February, March, and April of 1998, 1999, and 2000. This study shows that there are five key parameters, that is, surface rain rate, precipitation-top height (or temperature), and precipitation growth rates at upper, middle, and low layers to define a rainfall profile, and those five key parameters are strongly influenced by both SST and large-scale dynamics. Under the influence of 1997/98 El Niño, the precipitation-top heights in the EP were systematically higher by about 1 km than those under non–El Niño conditions, while the freezing level was about 0.5 km higher. Under the constraints of rain type, surface rain rate, and the precipitation top, the shape of rainfall profile still showed significant differences: the rain growth was relatively faster in the mid-layer (−5° to +2°C isotherm) but slower in the lower layer (below +2°C isotherm) under the influence of El Niño. It is also evident that the dependence of precipitation top height on SST was stronger under large-scale decent (non–El Niño) circulations but much weaker under large-scale ascent (El Niño) circulations. The combined effect of larger vertical extent and greater growth rate in the middle layer further shifted latent heating upward as compared with the impact of horizontal changes in the rain type fractions (convective versus stratiform). Such additional latent heating shift would certainly further elevate circulation centers and strengthen the upper-layer circulation.

Corresponding author address: Qilong Min, Atmospheric Sciences Research Center, State University of New York at Albany, 1400 Washington Avenue, Albany, NY 12222. E-mail: min@asrc.cestm.albany.edu

1. Introduction

Warming earth’s surface and lower troposphere associated with increasing greenhouse gases is likely to result in a more vigorous hydrologic cycle (Solomon et al. 2007). Because global warming is a relatively slow process (less than one degree per hundred years) and current global precipitation records are relatively short (several tens of years), accurately estimating the response of global precipitation to global warming remains a challenging problem (Solomon et al. 2007; Lambert et al. 2008). Observational evidence shows that there is significant disagreement of the response of global precipitation to global warming between observations and GCMs predictions (Soden 2000; Wentz et al. 2007). Although the real global warming pattern may be different from that of El Niño, the rapid warming of sea surface temperature (SST) in the east equatorial Pacific during El Niño events provides an ideal test bed to study the physical mechanisms of interaction between surface warming and precipitation (Soden 2000). Substantial efforts have been devoted to study the impacts of El Niño on precipitation (Ropelewski and Halpert 1987; Yulaeva and Wallace 1994; Dai and Wigley 2000; Curtis and Adler 2000; Adler et al. 2003; Chen et al. 2007; and references therein). Most of them, however, focused on horizontal features of cloud and precipitation systems, based on satellite imagery. Del Genio and Kovari (2002) found that storms over warmer oceans precipitate more heavily with larger horizontal storm size and higher cloud tops, but they do not have noticeably higher albedos than storms over cooler ocean waters. Schumacher and Houze (2003) pointed out that the stratiform rain fraction (SRF) dramatically increased in the central and east Pacific during the 1997/98 El Niño, indicating associated changes in the vertical distribution of latent heating. Additionally, Lin et al. (2006) also found horizontal cloud coverage associated with tropical deep convective system increases with SST in the warm tropical ocean area.

Response of tropical rainfall and cloud vertical structure to El Niño is receiving increasing attention in the recent years. Cess et al. (2001) and Wang et al. (2003) found that the occurrence of both high cumulative clouds and high subvisible ice clouds increased over the east Pacific (EP) during the 1997/98 El Niño. Berg et al. (2002) found that, in normal years, storm systems in the east Pacific were generally shallower and contained less ice particles compared with their counterparts in the west Pacific (WP). However, during the 1997/98 El Niño, rainfall vertical structures in these two areas became very similar. Li et al. (2005) pointed out the shape of rainfall profiles at given surface rain rate was changed in the east and central Pacific during the 1997/98 El Niño, accompanied by significant weakening of Walker circulation.

The precipitation variation at regional scale primarily reflects altered patterns of moisture transportation as well as local changes in evaporation (Trenberth 1998). Both dynamical and microphysical processes can also affect the rainfall vertical structure (Min et al. 2009; Li and Min 2010). Macrophysical features, including the precipitation-top height, the melting layer height, and the surface rain rate, are largely determined by dynamic conditions. The shape of rainfall profiles under the macrophysical constraints may be more related to microphysical processes. Furthermore, the change in atmospheric temperature is primarily determined by the residual between latent heating and radiative cooling (Graham 1995). Changes in rainfall profiles would result in shifts in latent heat profile and dynamical response. Using idealized latent heat profiles, Schumacher et al. (2004) demonstrated that the increase of stratiform rain fraction results in latent heat shifting to relatively higher altitudes, and this shift can elevate circulation centers and strengthen the upper-layer circulation. Therefore, it is crucial to understand the physical links of microphysical processes with dynamics in terms of rainfall and cloud vertical structures.

In this study we attempt to understand the El Niño–induced changes in rainfall vertical structure in the east Pacific, using Tropical Rainfall Measuring Mission (TRMM) precipitation radar (PR) observations and reanalysis data. We further investigate possible links of observed changes in rainfall vertical structure to various dynamical factors and their implications on latent heat profiles.

2. Data source

Investigating the vertical structure of precipitation is crucial for understanding dynamic and microphysical processes in cloud systems. The TRMM PR provides profile measurements of precipitation-sized hydrometeors at 4.3-km horizontal and 250-m vertical resolutions at nadir. We used the PR standard product 2A25 (Iguchi et al. 2000) to investigate the evolution of precipitation-sized hydrometeors under the influence of El Niño. The radar reflectivity represents the combined effect of hydrometeor effective size and total amount. In the TRMM PR 2A25 product, the young, active, and violent convection-related rain is identified as convective rain based on their very strong radar reflectivities, while the older, inactive, and weak convection-related rain is identified as stratiform rain based on their weak radar reflectivity and the feature of radar bright band (Steiner et al. 1995; Awaka et al. 1998; Houze 1997).

TRMM satellite was boosted to a higher orbit (400 km) in August 2001, resulting in some differences between pre- and postboost periods (DeMoss and Bowman 2007). Hence, we analyzed only the 1997/98 El Niño event and excluded the weak 2002/03 El Niño. We selected the region of 10°S–10°N, 120°–90°W in the EP to investigate El Niño–induced changes in precipitation and used the region of 10°S–10°N, 150°E–180° in the WP as a reference. We focused on the time period of January, February, March, and April 1998 (hereafter 1998JFMA), the later period of the 1997/98 El Niño event. Although there was a weak La Niña event that occurred in 1999, we found both the rainfall horizontal distribution and rainfall vertical structure during January–April 1999 were very similar to those during January–April 2000 (Li et al. 2005). Therefore, to minimize the weak La Niña effects, we combined the datasets for both January–April 1999 and January–April 2000 as a normal, basic state (hereafter, 9900JFMA). In addition to the above datasets, we also used the PR level-3 product of 3A25 and the Global Precipitation Climatology Project (GPCP) to provide monthly mean precipitation information.

TRMM PR-observed rainfall profiles generally use geometric altitude as the vertical coordinate. However, precipitation processes are directly associated with thermodynamic structure, that is, temperature, particularly for the microphysical processes. The relationship between the altitude and the temperature does change with time, particularly in the EP during an El Niño event. For example, the freezing level in EP was about 0.5 km higher in early 1998 than in early 1999. To better understand the precipitation (microphysical) processes, we converted the altitude-based rainfall profile to the air temperature–based rainfall profile. To do so, we collocated each TRMM PR observed rainfall profile with the nearest 6-hourly 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40) data at a resolution of 1.406° × 1.400° (256 × 128 regular Gaussian grid), based on the spatial nearest-neighbor method. Furthermore, to study the relationship between precipitation vertical structure and the large-scale circulation, we selected vertical velocity, relative humidity, and horizontal divergence, at low (850 hPa), middle (500 hPa), and high (200 hPa) levels, as key indicators; this information is also derived from the collocated nearest 6-hourly ERA-40 reanalysis data. Sea surface temperature (SST) has substantial influence on large-scale circulation and precipitation. To study the relationship between precipitation vertical structure and SST, the daily mean National Oceanic and Atmospheric Administration (NOAA) optimum interpolation sea surface temperature version 2 (OISSTv2) (Reynolds et al. 2002) in 1° × 1° resolution were assigned to each TRMM PR rainy pixel based on the nearest-neighbor method.

3. Results

a. The 1997/98 El Niño

The 1997/98 El Niño event began in the spring months of 1997 and ended in about May 1998. The SST in EP in 1998JFMA increased to about 30°C, which is about 3°C warmer than the climatology mean from 1982 (Fig. 1a). In the same time, the SST in WP slightly decreased. The selected EP area covers the warmest SST with the largest SST anomalies. The area mean air temperature profile shows that there was a systematic warming from sea surface to 200 hPa in 1998JFMA, as compared to the averaged profile in 9900JFMA. As reported by previous studies, the freezing level in 1998JFMA was ~0.5 km higher than that in 9900JFMA (Shin et al. 2000). The air temperature in WP, however, shows no significant change between the two periods.

Fig. 1.
Fig. 1.

Mean and anomalies of (a) sea surface temperature, (b) vertical velocity (Omega) at 500 hPa, and (c) RH at 500 hPa in tropical Pacific in 1998JFMA. Two red boxes represent the selected east and west Pacific areas. Accompanied with associated area mean vertical profiles in 1998JFMA (red curves) and 9900JFMA (blue curves).

Citation: Journal of Climate 24, 24; 10.1175/JCLI-D-11-00002.1

Significant changes in large-scale circulation due to the El Niño were also found in EP. Under the normal conditions, that is, 9900JFMA, the large-scale circulation was generally descent above 700 hPa with ascent at lower levels. During the El Niño event (1998JFMA), the low-level ascent was strengthened and the sign of vertical velocity of mid-high levels was reversed from descent to ascent. In WP, the maximum vertical velocity did not change but the corresponding altitude was shifted upward from about 700 to 500 hPa. As shown in Fig. 1b, our selected domain of EP covers the strongest updraft omega with the largest anomalies.

In addition, the atmospheric moistures also exhibited significant changes, as shown in Fig. 1c. The relative humidity (RH) profiles were similar below the ~600-hPa level. However, the RHs in layers above 500 hPa were much higher (10%–20%) in 1998JFMA than in 9900JFMA. The enhanced moisture due to excessive evaporation of warming SST was transported into the mid-upper layers by the ascent large-scale circulation under El Niño conditions. In contrast, the RH profiles in WP showed no significant changes under El Niño conditions.

Storm systems respond to those significant changes in dynamics and thermodynamics, resulting in changes in both horizontal and vertical structures of precipitation. Before studying specific links of precipitation properties with dynamics and thermodynamics in our selected EP and WP areas, we first used the PR 3A25 product to illustrate the macrophysical characteristics of rainfall three-dimensional structures in response to El Niño in the Pacific basin. Figures 2 and 3 show the comparison of three rainfall properties for convective and stratiform rain between 1998JFMA and 9900JFMA, respectively. The selected parameters are the near-surface rain rate at 2 km, the precipitation-top height, and the relative growth rate of rain rate between 4 and 6 km normalized by the near-surface rain rate.

Fig. 2.
Fig. 2.

Comparison of monthly mean convective rain properties in tropical Pacific in 1998JFMA with those in 9900JFMA: (a),(d) the near-surface rain rate; (b),(e) the storm height; and (c),(f) the ratio of difference between rain rate at 6 and 4 km [CR(4km)-CR(6km)] to rain rate at 2 km [CR(2km)], where CR stands for convective rains.

Citation: Journal of Climate 24, 24; 10.1175/JCLI-D-11-00002.1

Fig. 3.
Fig. 3.

As in Fig. 2, but for stratiform rain.

Citation: Journal of Climate 24, 24; 10.1175/JCLI-D-11-00002.1

As shown in Fig. 2, during 1998JFMA, the Pacific intertropical convergence zone (ITCZ) shifted to the south and the Southern Pacific convergence zone (SPCZ) moved northward and joined with the ITCZ to form a single convergence zone. During 9900JFMA under non–El Niño conditions, the ITCZ and SPCZ were clearly separated with a distinguished double-ITCZ structure in EP (Zhang 2001). The mean near-surface rain rates of both convective and stratiform rains in EP were relatively heavier during 1998JFMA than during 9900JFMA. In addition to the changes of surface rain distribution, the stratiform rain fractions in EP significantly increased in 1998JFMA (Schumacher and Houze 2003). Those changes of horizontal characteristics of precipitation during an El Niño event were well documented (Ropelewski and Halpert 1987; Yulaeva and Wallace 1994; Dai and Wigley 2000; Curtis and Adler 2000; Adler et al. 2003; Chen et al. 2007; and references therein).

The rainfall vertical structures in EP also changed significantly during the El Niño event as shown in both Figs. 2 and 3. The mean precipitation-top heights (hereafter PTH) in EP during 1998JFMA were obviously higher than those during 9900JFMA. On the other hand, the PTHs in WP did not change substantially, although the rainfall horizontal distributions were quite different. It indicates the changes in rainfall vertical structure in WP were relatively small. Because the probability distribution function of PTH was a bimodal pattern (see section 3b), it is inadequate to use the monthly mean value to investigate the detailed PTH changes under the influence of El Niño. To reveal the physical implication of SST and large-scale circulation on the PTH, we must analyze the detailed rainfall profiles from orbital TRMM PR measurements.

In addition to the changes in PTH, the relative rain growth rate at certain layer changed as well. The TRMM PR 3A25 product provides the rain rates at three altitudes of 2, 4 and 6 km. As the fastest growth of rain occurs between 4 and 6 km (Liu and Fu 2001), we used a normalized growth rate of rain rate, i.e., , as an indicator to represent the rain growth process. This index indicates the fraction of near surface rain that is formed from the 4–6 km layer. As shown in Figs. 2c, 2f, 3c, and 3f, the growth rates of rain rate were higher in EP during 1998JFMA than during 9900JFMA, especially for the stratiform rains.

Above analyses clearly show that El Niño has significant impacts on both horizontal and vertical distributions of rainfall in pacific basin. We will focus on the changes of precipitation vertical structure under the influence of El Niño since less attention has been paid. Specifically, we will investigate the response of precipitation-top height (PTH) and rain growth rates at different vertical layers to El Niño, in terms of SST and large-scale dynamics, by using TRMM PR rainfall profile product 2A25.

b. Changes in precipitation top

PTH is one of the most important parameters of rainfall vertical features. The PTH directly reflects how well a storm develops vertically and is strongly determined by thermal and dynamic conditions. In this study, we defined the PTH of each rainfall profile in the product 2A25 as the highest altitude with at least three continuous radar bins with rain echo. Further, we converted PTH to precipitation-top temperature (PTT), by collocating the profile with the nearest 6-hourly ERA-40 atmospheric state.

Without separation of rain type, Fig. 4 shows the overall PTH probability distribution functions (PDFs) in the selected EP and WP under both El Niño and non–El Niño conditions. PTHs in EP and WP show a bimodal distribution with the low mode at 2 ~ 4 km and the high mode peaked at 6 ~ 7 km, consisting with the global tropical PTH distributions (Short and Nakamura 2000). The bimodal distribution indicates the partition of shallow convections and deep convections. Under non–El Niño conditions, the maximum occurrence of the low mode in EP was comparable to that of the high mode, but much larger than the maximum occurrence of the low mode in WP. Furthermore, the altitude associated with the maximum occurrence of the low mode in EP was relatively lower (close to 2km) than that in WP. It suggests that much shallower convections prevailed in EP under non–El Niño conditions.

Fig. 4.
Fig. 4.

PDF of storm height in the selected (a) east and (b) west Pacific areas in 1998JFMA and 9900JFMA. All types of rain detected by TRMM PR are included.

Citation: Journal of Climate 24, 24; 10.1175/JCLI-D-11-00002.1

On the other hand, the occurrence of the high mode in EP under non–El Niño conditions was lower than that in WP. The large-scale circulation in WP, driven by warm SST, was prevalently upward, while the vertical motion in EP, driven by SST gradient, was sometime downward (Back and Bretherton 2006). As deep convection originates preferentially in environments with upward motion, this explains the observed contrast in PTH distribution between WP and EP. Similar features of cloud-top distribution were also observed by CloudSat (Kubar and Hartmann 2008). However, under El Niño conditions, the SST in EP became equal to or even higher than that in WP. The deep convections in EP were enhanced and the shallow convections were suppressed. As shown in Fig. 4a, both occurrence and mean PTH of the high mode of the PTH distribution in EP under El Niño conditions were larger than those under non–El Niño conditions. In the meantime, the PTH distribution in WP between El Niño and non–El Niño conditions showed no significant difference (Fig. 4b).

In general, PTH increased with the surface rain rate, as shown in Fig. 5, for both convective and stratiform rains under El Niño and non–El Niño conditions. PTHs under El Niño conditions were systematically higher by about 1 km than those under non-El Niño conditions. It implies that for given column integrated latent heat (or surface rain rate), latent heat release may occur higher in the atmosphere in EP under El Niño conditions.

Fig. 5.
Fig. 5.

Mean relationship between precipitation-top height and the near-surface rain rate in EP observed by TRMM PR.

Citation: Journal of Climate 24, 24; 10.1175/JCLI-D-11-00002.1

Precipitation and associated clouds are affected by local SST changes through the activation of thermodynamical processes within the atmospheric column. Precipitation is also sensitive to many other factors that may depend only partly on the local SST, resulting in a complicated dependence of PTH on SST, shown in Fig. 6. With all samples within 1998–2000 JFMA, there was a statistically significant correlation (the correlation coefficient R = 0.19, the probability of being false P < 0.000 01) between PTH and SST, with a slope (or sensitivity) of 0.37 km (5.71%) °C−1 for convective rains. However, if separating the El Niño data samples from the non–El Niño samples, there were two distinct dependences of PTH on SST: 1) a weak correlation (R = 0.04, P = 0.07) with a low sensitivity of 0.13 km °C−1 under El Niño conditions; and 2) a relatively stronger correlation (R = 0.09, P < 0.000 01) with a still low sensitivity of 0.20 km °C−1 under non–El Niño conditions. The overall high sensitivity and stronger correlation of PTH with SST for all data samples were mainly due to a significant increase of mean PTH (~1 km, shown in Fig. 5) under El Niño conditions. SST variations are very inhomogeneously distributed over the tropical ocean and are thus associated with profound modifications of the large-scale atmospheric circulation. As discussed previously, due to SST warming of El Niño, the large-scale circulation in EP reversed from descent to ascent at midhigh levels with a strengthened low-level ascent. In WP, as large-scale circulation patterns were persistent from 1998 to 2000, the relationship between PTH and SST during the 1998JFMA (R = 0.16) was also quite consistent with that of during 9900JFMA (R = 0.19). Observed dependence of PTH on SST thus partly stems from changes in large-scale dynamics.

Fig. 6.
Fig. 6.

Scatterplots of SST to PTH for (a) convective rain and (b) stratiform rain. Dark gray cross for 1998 JFMA; light gray cross for 9900JFMA. The associated mean relationships (0.5°C SST interval) with standard deviation for 1998JFMA and 9900JFMA are overlapped. Additionally, the associated linear regression results are also overlapped: the gray lines stand for 1998JFMA, dark gray lines for 9900JFMA, and black thick lines stand for the whole 12 months.

Citation: Journal of Climate 24, 24; 10.1175/JCLI-D-11-00002.1

Since the freezing level in EP was about 0.5 km higher in early 1998 than in early 1999 and 2000, it is interesting to know if observed changes of PTH are simply due to thermoexpansion of the atmosphere in response to SST warming or other fundamental dynamical changes. As listed in Table 2, the PTT was also dependent on SST, with a relatively weaker correlation and smaller sensitivity than those of PTH. It suggests that, although a portion of precipitation-top growth are due to the thermo-expansion, the observed dependence of PTH (or PTT) on SST indeed stems from changes in large-scale dynamics associated with SST warming.

As the large-scale atmospheric circulation transports heat and moisture and affects the thermodynamical stability of the atmosphere, it is crucial to better understand associated large-scale dynamic and thermodynamic changes and its impacts on PTH (or PTT). Using vertical velocity, relative humidity, and horizontal divergence, at low (850 hPa), middle (500 hPa), and high (200 hPa) levels, as key thermodynamic and dynamical indicators, we investigated the correlations and sensitivities of precipitation top (PTH and PTT) to them. Direct measurements of those parameters at cloud scale from satellite measurements are not available. Those dynamical indicators inferred from reanalysis (ERA-40) represent large-scale dynamics of ambient air around precipitation and associated clouds. The statistical results are summarized in Table 1 (for PTH) and Table 2 (for PTT), respectively.

Table 1.

The correlation coefficient (R) and the sensitivity (S) between PTH and SST and circulation parameters. In each cell, the first value is for the whole 3 years. The value in parenthesis is for 1998 (El Niño), the value in square brackets is for 1999 and 2000 (non–El Niño). Results that passed the F-test 99% confidence level are in bold, and results that failed the 99% test but passed the 90% confidence are roman. Those that failed the 90% test are marked as “N/A.”

Table 1.
Table 2.

As in Table 1, but between PTT and SST and circulation parameters.

Table 2.

Overall, all those indicators were statistically significantly correlated with precipitation-top height (and temperature) for both convective and stratiform rains if all samples (El Niño and non-El Niño) were included. In general, except at 850 hPa, the PTH (and PTT) of stratiform rains had stronger correlations with the ambient dynamic and thermodynamic conditions than that of convective rains. It is consistent with our understanding that the stratiform rains are more dependent on large-scale environmental factors while convective rains are mainly controlled by dynamics at local and convectional scales.

As discussed previously, such statistically significant correlations and sensitivities (i.e., the linear regression slope of PTH or PTT to each of selected parameters) stem from changes of large-scale circulation due to the El Niño event. The interesting question is how precipitation (and cloud in general) responds to those dynamical indicators with respect to a similar large-scale circulation pattern, either El Niño or non–El Niño conditions. Under El Niño conditions, PTH of convective rains had stronger correlation (R) and larger sensitivity (S) with vertical velocity of the upper layer at 200 hPa than with those of the lower layers at 500 and 850 hPa, However, under non–El Niño conditions, the convective rain PTH had the highest correlation coefficient with vertical velocity of the middle layer at 500 hPa, although the maximum sensitivity was still with the upper layer at 200 hPa. Those characteristics were also true for stratiform rains and true for PTT. Certainly, stronger upward motion favors the vertical expansion of storm systems to lift precipitation top higher. The fact that the strongest correlation between PTH and vertical velocity occurred at the layer above the mean precipitation top (about 400 hPa for 1998JFMA and 550 hPa for 9900JFMA) may also suggest that observed correlation may be, in part, due to the ambient atmospheric response to the upward shift of latent heating of precipitation. Thus the concurrence of relatively stronger ascent at 200 hPa and the higher PTH (or colder PTT) in both convective and stratiform rains, indicates a possible dynamical interaction between precipitation top and the upper-layer circulation. Response of large-scale circulation to the shift of latent heat has been studied by Mapes and Houze (1995) and Schumacher et al. (2004). However, in their study, the elevated latent heat distribution is purely due to the increase of stratiform rain fraction in response to El Niño. What is observed here is that the vertical stretch of precipitation redistributes the latent heat to higher layers (more discussion in section 3d).

The vertical development of convections also depends on the RH in the ambient air. The entrainment of the relative dry environmental air can evaporate the condensate. In general, the more moisture in the ambient air, the higher PTH and colder PTT were. Under El Niño conditions, PTH had relatively larger correlation coefficients with the mid-upper layer RH at 500 hPa than with the low layer RH at 850 hPa with both maximums of correlation coefficient and sensitivity, particularly in the stratiform rain regime (no significant correlation at 850 hPa). Although the boundary layer moisture mainly determines the column precipitable water and consequently the surface rain amount, moist midlevel humidities are most important to the vertical development of storm, which favor penetration of convective elements to the upper troposphere with larger vertical extent and higher PTH. Certainly, evaporation of precipitating hydrometeors would result in the enhancement of moisture, particularly in stratiform rain region.

Beside the vertical velocity, the horizontal divergence is also an important indicator of large-scale circulation, although they have strong inherent correlations. Strong divergence at the upper level of 200 hPa generally associates with strong updraft (i.e., negative Omega), and strong convergence (negative divergence) at the low level of 850 hPa associates with strong updraft. For all cases, PTH (PTT) was higher (colder) with increasing divergence at 200 hPa, consisting with the correlation between PTH (PTT) and the vertical velocity. Under El Niño conditions, however, PTH increased with the convergence (i.e., negative divergence) at the mid-layer of 500 hPa. Given both maximum of correlation coefficient and sensitivity between PTH and RH at this midlevel of 500 hPa, the positive correlation between PTH and convergence suggests that the mid-layer around 500 hPa is a “sink” of moisture that condensates there and feeds the convections for the vertical development of storm system and precipitation. More discussion will be given later that the mid-layer around 500 hPa is also a key region for rain growth. On the other hand, PTH showed no correlation with the low-level divergence under El Niño conditions. In contrast, under non–El Niño conditions, PTHs for both convective and stratiform rains were strongly correlated with the low-level convergence. As a strong SST gradient should result in a strong low-level convergence, we speculate that storms in EP under non–El Niño conditions are driven by horizontal SST gradient while storms under El Niño conditions are driven by warm SST through thermal convection.

Variations in the large-scale dynamics affect clouds and precipitation, and since the dynamics is also tied to the SST pattern, further analysis to deconvolve those factors is necessary and crucial. Figure 7 shows distributions of the mean PTT for given SST and given vertical velocity, RH, or divergence at 200 hPa. Although it is somewhat noisy due to limited samples, it clearly shows that for given dynamic conditions, PTT decreased generally with SST when SST was colder than ~26°C, that is, within the SST range under non–El Niño conditions, and PTT showed no significant change at warmer SST range under El Niño conditions. For a given SST, PTT decreased with vertical velocity, relative humidity, and divergence at 200 hPa, respectively. PTT also decreased with relative humidity and convergence at 500 hPa, respectively, and had insignificant change with other indicators at other levels (not shown). It is consistent with our previous analyses.

Fig. 7.
Fig. 7.

Mean PTT under given SST and given (a),(d) Omega, (b),(e) RH, and (c),(f) the divergence at 200 mb for (a)–(c) convective and (d)–(f) stratiform rains in 1998–2000 JFMA in EP.

Citation: Journal of Climate 24, 24; 10.1175/JCLI-D-11-00002.1

c. Changes in rainfall vertical profiles

The TRMM PR measures instantaneous rainfall profiles. The vertical gradient of the rainfall profile is related to microphysical processes of raindrop growth and thermodynamical processes of the storm system. The overall mean rainfall profiles in EP in El Niño and non–El Niño (shown in Fig. 8a) have been studied by Berg et al. (2002). To understand the changes of the “relative” shape of rainfall profiles under the influence of El Niño, dynamic processes must be constrained. Various dynamical constraints have been used in previous studies, that is, rain type (convective–stratiform–shallow) and surface rain rate (Rsrf, Liu and Fu 2001; Li et al. 2005). As discussed in the previous section, the precipitation top (PTH or PTT) is also one critical factor that is directly related to ambient dynamical parameters and Rsrf. Therefore, we used both PTT and Rsrf as dynamic constraints, in additional to rain type.

Fig. 8.
Fig. 8.

(a) Overall mean rain rate profiles, (b) mean profiles for convective rains, and (c) for stratiform rains under given PTT and Rsrf in EP.

Citation: Journal of Climate 24, 24; 10.1175/JCLI-D-11-00002.1

Figures 8b and 8c show examples of the mean rainfall profile under both El Niño and non–El Niño with fixed PTT (−40° to −20°C and −20° to 0°C) and Rsrf, for different rain types. It is clear that even under those strict constraints, the shape of rainfall profiles under El Niño conditions can be different from those under non–El Niño conditions for various rain types.

The water vapor condensation–sublimation mainly occurs in the cloud formation processes rather than in the rain growing processes. As described by Kessler’s auto-conversion theory, the rain growth rate depends on both rain rate and cloud water content. There were no global observations of cloud water content profile, particularly directly associated with rain rate during the 1997/98 El Niño. The vertical and horizontal advections result in displacement between where rain grows and where condensation and evaporation happen. However, the rainfall profile is the ultimate outcome of rain forming and growth processes. Thus it directly links to the rain growth processes. Based on the time–space exchangeable assumption and ignoring horizontal advection, the vertical gradient of the rainfall profile can be expressed as
e1
where R is the rain rate; Z is the height; and ϖ are the mean falling speed and the mean updraft velocity, respectively; and W is the rainwater content of precipitation hydrometeors. If assuming the invariance of the net fall speed at the instance of PR measurement, we have
e2
hence the vertical gradient of the rainfall profile approximately represents the net formed hydrometeors at each layer. It is important to emphasize that the above derivation is under a lot of assumptions and some assumptions may not hold in each individual rainfall profile. In a statistical sense, nonetheless, the vertical gradient of the rainfall profile can be viewed as the first-order approximation of the rain growth rate and associated latent heat rate, which has been exploited by many researchers (Rogers and Yau 1989, p. 131; Tao et al. 1990; Yang and Smith 1999; Liu and Fu 2001; Satoh and Noda 2001).

Liu and Fu (2001) pointed out that TRMM PR-observed altitude-based rainfall profiles can generally be parameterized by several piecewise log-linear curves. This relationship is also valid for the temperature-based profiles (see Fig. 8a). Based on the vertical gradients of the rainfall profile and the differences of rainfall profile between El Niño and non–El Niño we divided rainfall profiles into three layers and defined the associated linearly regressed slopes of logR ~ T as the following: 1) SlopeA in the upper layer with temperature colder than −5°C; 2) SlopeB in the melting layer with temperature −5°C to +2°C; and 3) SlopeC in the lower layer with temperature warmer than +2°C. Since TRMM PR, a single-frequency radar, cannot distinguish the size and phase of precipitating hydrometeors, interpreting the observed changes of radar reflectivity as changes in precipitating water content is too simplistic, particularly in the stratified bright band near the freezing level in the stratiform rain region where hydrometeor phase transition occurs. All measurement errors and retrieval uncertainties have some impacts on the absolute values of rain rate profiles. However, our study is focused on the impacts of El Niño on rainfall profiles, that is, the contrast between El Niño and non–El Niño. Our defined slopes are used as an index to quantify changes in rainfall profile. With the same retrieval algorithm and processing procedure, measurement errors and retrieval uncertainties have limited impacts on using those slope indexes to characterize the rainfall profile changes, particularly in the difference between El Niño and non–El Niño.

In convective rains, rain rate tends to increase toward the cloud base over ocean, a maximum rain rate occurs at the 2–3-km altitude, and below that the rain rate tends to decrease toward the earth, due to either breakup and/or evaporation (Liu and Fu 2001). Therefore, we only used the values above the altitude of maximum rain rate to calculate the SlopeC. For those stratiform rains linked to deep convection system, the SlopeC is near to zero or even negative. However, some shallow–warm rains are also included in the category of the stratiform rains in TRMM PR 2A25 product. Those shallow–warm rains keep growing toward the earth even in low layers, resulting in the SlopeC of stratiform rain to be more positive. And such an effect is more obvious under non–El Niño conditions because of the relatively higher occurrence of warm rains. As discussed by Houze (1993) and Liu and Fu (2001), the SlopeA is small and dominated by the water vapor deposition and the riming processes. The SlopeB is the biggest among the three and represents the rapid growth of rain dominated by aggregation and riming. The SlopeC represents the warm rain process mainly growing through coalescence of cloud droplets in the convective rains.

The shape of rainfall profile, as represented by SlopeA, SlopeB, and SlopeC in Figs. 9a and 9b, depends on rain type, Rsrf, and PTT. Generally, for a given Rsrf, all slopes decreased with decreasing of PTT (i.e., increasing precipitation top) as the fixed surface rain amount was formed from a thicker precipitating layer. For a given PTT, SlopeB clearly increased with Rsrf, while SlopeA and SlopeC only weakly depended on Rsrf. It suggests that although the rain forming processes in the layers associated SlopeA and SlopeC contribute to the surface rain rate, the rain forming and growth processes in the layer +2 to −5°C are crucial to determine the surface rain rate.

Fig. 9.
Fig. 9.

(a) The mean SlopeA, SlopeB, and SlopeC as functions of surface rain rate and precipitation-top temperature in (left) 1998JFMA and in (right) 9900JFMA for convective rains. (b) As in (a), but for stratiform rains.

Citation: Journal of Climate 24, 24; 10.1175/JCLI-D-11-00002.1

Comparing those slopes under El Niño and non–El Niño conditions, we found no significant difference for SlopeA. However, even for given PTT and surface rain rate, the SlopeB (SlopeC) under El Niño conditions is greater (smaller) than that under non–El Niño conditions significantly. It indicates that under the same dynamic constraints the precipitating hydrometeors grow faster in the mixed phase layer but slower in the lower layer under El Niño conditions.

In the low layer, there was slightly higher growth rate, as indicated by SlopeC, under non–El Niño conditions than under El Niño conditions for convective rains. Some negative values of SlopeC (reduction rate of rainfall toward the earth’s surface) occurred in the layer, particularly for the stratiform rains, indicating raindrop evaporation and/or breaking up processes. Under non–El Niño conditions, the large descent ambient air in EP favored the warm rain processes, resulting in a relatively larger SlopeC. In contrast, under El Niño conditions, the large-scale circulation in EP turned to be ascent and transported the low-level moisture into middle and upper layers, and thus favored the mid-layer rain forming processes, resulting in relatively larger SlopeB and associated smaller SlopeC.

Statistical analysis of all slopes with SST and dynamical indicators confirms that the SlopeB is the most sensitive parameters to the changes of large-scale circulation associated with El Niño than SlopeA and SlopeC. Hence, we only listed correlation coefficients (R) and the sensitivities (S) of SlopeB with SST and circulation indicators in Table 3. Overall, the SlopeB of stratiform rains was more sensitive to those parameters than that of convective rains, as expected. For convective rains, the SlopeB did not significantly correlate with SST although its mean value under El Niño conditions (0.084) was remarkably larger than under non–El Niño conditions (0.075), indicating the fast growth of rain rate at the mixing layer was not directly controlled by the SST. Instead, the SlopeB was strongly correlated with the vertical velocity, the relative humidity, and the divergence of the upper level at 200 hPa, as well as with the relative humidity at 500 hPa under El Niño conditions and had no statistically significant correlation under non–El Niño conditions. It again suggests that the observed correlation may be partly due to the ambient atmospheric response to changes of latent heating of precipitation under the influence of El Niño. The fact that SlopeB under non–El Niño conditions, that is, the descent large-scale circulation, did not significantly correlate with the selected thermodynamic and dynamic parameters warrants further study.

Table 3.

The correlation coefficient (R) and the sensitivity (S) between SlopeB (PTT < −10°C) and SST and the circulation parameter. In each cell, the first value is for the whole 3 years. The value in parenthesis is for 1998 (El Niño), and the value in square brackets is for 1999 and 2000 (non–El Niño). Results that passed the F-test 99% confidence level of are in bold, and results that failed the 99% test but passed the 90% confidence are roman. Those that failed the 90% test are marked as “N/A”.

Table 3.

d. Implications to latent heat estimation

The results shown above indicate that even for a given geolocation, season, rain type, and surface rain rate, the vertical structure of precipitation can be significantly different under different large-scale dynamics (El Niño versus non–El Niño) and its response to SST warming is also different. Changes in precipitation vertical structure have important implications on latent heating distribution, and thus subsequently on large-scale circulation. To illustrate the importance of considering precipitation vertical distribution, we estimated latent heat (LH) distribution in EP domain using observed precipitation profiles.

As discussed in previous section, that is, Eq. (2), the vertical gradient of the rainfall profile is an indicator of the net formed hydrometeors at each layer. Therefore, we have
e3
where LH is the latent heating rate (k h−1), ρa the density of dry air, Cp the certain heat of water at constant pressure (J kg−1 k−1), fc–e is the mass fraction of condensation–evaporation, ff–m is the mass fraction of freezing–melting, fd–s is the mass fraction of deposition–sublimation; and Lυ, Lf, and Ls are the associated heats of water phase changing (condensation, freezing, and sublimation). Equation (3) illustrates that LH is directly links to the vertical gradient of the rainfall profile, particularly if further assuming the horizontal divergence of rain amount is averaged out over the large selected domain and the net amount of nonprecipitating clouds for long duration (monthly) is zero. To quantitatively estimate latent heat from the rain rate requires complicated partitions of microphysical properties of the hydrometeors. Precisely estimating LH profiles, however, is out of the scope of this paper. We directly used the vertical gradient of the rainfall profile as an index of LH profile.

Although the vertical distribution of rainfall profiles varies substantially within each rain type, we used the mean vertical distribution of normalized rainfall profile with respect to the Rsrf for each rain type to simplify our calculation. Figures 10a, 10b, and 10c show the mean normalized rainfall profile for convective rains, stratiform rains, and warm rains, respectively. The warm rains (with PTH lower than the freezing level) are separated from the convective and stratiform rains to eliminate the potential errors introduced by the normalization processes.

Fig. 10.
Fig. 10.

The mean rain rate profiles of (a) convective, (b) stratiform, and (c) warm rains normalized by Rsrf and (d)–(f) the associated vertical gradient profiles (times by 250 m).

Citation: Journal of Climate 24, 24; 10.1175/JCLI-D-11-00002.1

Figures 10d, 10e, and 10f show the associated profiles of vertical gradient of the normalized rain rate per TRMM PR vertical bin (250 m), representing the fraction of surface rain amount that is formed at the layer within thickness of 250 m. As expected, the vertically integrated value of the gradient profile equals to unity (or 100%). Clearly, more hydrometeors formed from higher altitude under El Niño conditions than under non–El Niño conditions, indicating upward shift of latent heat due to El Niño for all types of precipitation.

LH is a result of the phase changes of water in the atmosphere, involving cloud and rain forming and growth processes. The simplified LH index of dR/dZ neglects some contributions of LH associated with nonprecipitating clouds. Also a simplistic attribution of observed changes of radar reflectivity as changes in precipitating water content may introduce some uncertainties in LH estimates, particularly in the stratified bright band near the freezing level. However, the goal of this study is to demonstrate how the changes in rainfall profiles, for given rain types, can impact the LH vertical distribution and more importantly the difference of LH profiles between El Niño and non–El Niño conditions. The LH index of dR/dZ provides a direct linkage of rainfall profiles to LH profiles and represents a first-order approximation of LH profile. There is no existing LH product that directly utilizes the full information of TRMM PR-observed rainfall profiles. Therefore, we used the LH index of dR/dZ for our following discussion.

To estimate LH in EP during 1998JFMA, we combined the above vertical gradient of the normalized rain rate with precipitation statistics of mean surface rain amount, a fraction of three rain types. Table 4 lists the precipitation statistics of mean surface rain amount (from GPCP monthly data) and the fractions of three rain types (from TRMM PR 3A25). The mean surface rain amounts in EP were about 6.73 mm day−1 during 1998JFMA and 2.53 mm day−1 during 9900JFMA, respectively. The stratiform rain fraction during 1998JFMA (46.06%) were 16% higher than that during 9900JFMA (29.08%). The warm rain fraction were ~11% lower during 1998JFMA (12.23%) than during 9900JFMA (22.93%).

Table 4.

Statistics of mean rain amount, the rain fraction of convective, stratiform, and warm rains in the selected eastern Pacific area.

Table 4.

There are three different ways to estimate the domain-averaged LH in EP during 1998JFMA: 1) Test 1 (T1) is to use the climatologic mean vertical gradient of the normalized rain rate, that is, the mean profile of 9900JFMA averaged over three rain types with corresponding rain type fraction of 9900JFMA, and scale it with observed surface rain amount during 1998JFMA; 2) Test 2 (T2) is to use the climatologic vertical gradients of the normalized rain rate of 9900JFMA for each rain type and scale them with observed rain type fraction and surface rain amount during 1998JFMA; and 3) Test 3 (T3) is to use observed vertical gradients of the normalized rain rate of 1998JFMA for each rain type and scale them with observed rain type fraction and surface rain amount during 1998JFMA. T1 is a simple and common practice used in the literature. T2 is proposed by Schumacher and Houze (2003) and addressed the importance of stratiform fraction changes and its consequence on large-scale circulation (Schumacher et al. 2004). T3 is designed to address the impacts of both stratiform fraction and vertical structure changes on the LH distribution.

Stratiform rains heat the middle–upper atmosphere and cool the bottom of the atmosphere. With an increase of stratiform rain fraction during 1998JFMA, LH vertical distribution shifts upward, resulting in positive differences above 4 km and weakly negative differences below. The estimated profiles of LH index in 1998JFMA in EP using T1, T2, and T3 methods and the associate differences among them are shown in Fig. 11. The maximum difference between T2 and T1 is up to 5% at about 4.5 km. It is qualitatively consistent with the finding of Schumacher et al. (2004). Since the precipitation top elevated and the midlevel rain growth rate increased during 1998JFMA for a given surface rain rate for all three rain types, the estimated LH of T3 shifts further upward, in addition to the impact of the increase of stratiform fraction (T2). There are strong positive differences (up to 15%) above ~5 km and strong negative differences (up to −20%) below, as compared with T2. Comparing all three estimates of LH profiles, the impact of the rainfall vertical structure is at least comparable, if not much stronger, to the impact of the rain type fractions. Such additional latent heating shifts would certainly have substantial impacts on large-scale circulation, particularly on the upper atmosphere. It is worthy to note that the quantitative statistics of the rain type fraction and associated mean normalized rainfall profiles may vary with different classification criteria for each rain type (Schumacher and Houze 2003). However, using slightly different classification criteria would not alter our conclusion.

Fig. 11.
Fig. 11.

(a) The estimated profiles of LH index in 1998JFMA in EP based on three tests (T1, T2, T3). (b) The relative difference among those three results.

Citation: Journal of Climate 24, 24; 10.1175/JCLI-D-11-00002.1

4. Conclusions

The rapid warming of sea surface temperature in the east equatorial Pacific during El Niño events provides an ideal test bed to study the physical mechanisms of interaction between surface warming and precipitation. To understand the physical links of rainfall growth with dynamics, we investigated the 1997/98 El Niño–induced changes in rainfall vertical structure in the east Pacific, using collocated TRMM PR and associated daily SST and 6-hourly reanalysis data during JFMA of 1998, 1999, and 2000. We segregated PR measurements into convective and stratiform rain categories and compared their statistical characteristics under El Niño conditions (1998JFMA) with those under non–El Niño conditions (9900JFMA), and linked the changes of precipitation vertical structure to SST and vertical velocity, RH, and divergence at different vertical levels.

To describe rainfall profiles for both convective and stratiform rains, we proposed five key parameters of surface rain rate, precipitation-top height (or temperature), and precipitation growth rates at upper, middle, and low layers. The first two are mainly controlled by dynamical processes of storm systems. The latter three precipitation growth rates at different layers correspond to different microphysical mechanisms of precipitation forming and growing, that is, the upper level (cold rain or ice rain) by slow water vapor deposition, the middle level (melting layer or mixed phase rain) by rapid aggregation and riming, and the low level (warm rain) by coalescence, respectively. Each precipitation regime, that is, convective rain and stratiform rain, exhibits its own characteristics. Our study showed that the five key parameters of rainfall profile are strongly influenced by both SST and large-scale dynamics. It is worthy to notice that all measurement errors and retrieval uncertainties have some impacts on the absolute values of rain rate profiles. However, our study is focused on the impacts of El Niño on rainfall profiles, that is, the contrast between El Niño and non–El Niño. With the same retrieval algorithm and processing procedure, measurement errors and retrieval uncertainties have limited impacts on using key parameters to characterize the rainfall profile changes, particularly in the difference between El Niño and non–El Niño.

Under the influence of 1997/98 El Niño, PTHs in EP were systematically higher by about 1 km than those under normal non–El Niño conditions, while the freezing level was about 0.5 km higher. Overall, there was a strong correlation (R = 0.19, P < 0.000 01) between PTH and SST, with a sensitivity (or slope) of 0.37 km (5.71%) °C−1 for convective rains during the 1998–2000 JFMA. However, under El Niño ascent mid-upper circulations (1998JFMA), the sensitivity was about at 0.13 km °C−1 (one-third of overall value) with a weak correlation (R = 0.04, P = 0.07). In contrast, under non–El Niño decent circulations (9900JFMA), the sensitivity increased to 0.20 km °C−1 with a relatively strong correlation (R = 0.09, P < 0.000 01). Many other rainfall profile parameters also exhibited similar characteristics from our study. It is evident that the dependence of vertical structure of precipitation on SST is strongly sensitive to large-scale dynamics: a large response of rainfall profile to SST under large-scale decent circulations and a small response under large-scale ascent circulations. Therefore, the observed rainfall structure changes cannot be attributed simply to the local SST warming because large-scale dynamics plays a key role in determining rainfall structure and the large-scale dynamics is only partly tied to the SST pattern. Hence, inferring cloud and precipitation feedback of climate warning by simply examining SST dependence alone could draw a misleading conclusion.

The boundary layer moisture is important in determining the column precipitable water and the surface rain rate. However, moist midlevel humidities are most important to the vertical development of a storm. Even under the same constraints of rain type, surface rain rate, and the precipitation top, the shape of rainfall profile still showed significant differences: the rain growth is relatively faster in the mid-layer (−5 to +2°C isotherm) but slower in the lower layer (below +2°C isotherm) under the influence of El Niño. This stems from the changes of large-scale circulation in EP due to El Niño, which turns out to be ascent and transports the low-level moisture into middle and upper layers, favors the midlayer rain-forming processes.

The concurrence of relatively stronger ascent at 200 hPa and the higher PTH (or colder PTT) in both convective and stratiform rains, thus, indicates a possible dynamical interaction between precipitation top and the upper-layer circulation. On one hand, stronger upward motion favors the vertical expansion of storm systems to lift precipitation top higher. On the other hand, the upper-atmosphere responses to the upward shift of latent heating released from precipitation, as a result of rainfall vertical structure changes, to elevate circulation centers and strengthen the upper-layer circulation. Changes in rain profiles result in changes in latent heat profiles and consequently in atmospheric dynamical response to SST warming or El Niño. The combined effect of larger vertical extent and greater growth rate in the middle layer, found in this study, could result in substantial changes in latent heat profile.

We further used the vertical gradient of rainfall profile dR/dZ as an index of LH to evaluate the impacts of changes of rainfall profiles on LH vertical distribution.

Although the LH index of dR/dZ is very simplistic, it provides a direct linkage of rainfall profile to LH vertical distribution and represents a first-order approximation of LH profile. As demonstrated through three test scenarios, there are strong positive differences (up to 15%) above ~5 km and strong negative differences (up to −20%) in addition to changes associated with horizontal redistribution of convective rain and straitform rain under El Niño conditions reported by Schumacher et al. (2004). Comparing all three estimates of LH profiles, the impact of the rainfall vertical structure is at least comparable, if not much stronger, to the impact of the rain type fractions. Such additional latent heating shifts would certainly further elevate circulation centers and strengthen the upper-layer circulation. Therefore, it is crucial to consider changes in the vertical distribution of precipitation in model evaluation and in LH retrieval algorithm development.

Acknowledgments

This research was supported by the Office of Science (BER), U.S. Department of Energy, Grant DE-FG02-03ER63531, the NOAA Educational Partnership Program with Minority Serving Institutions (EPP/MSI) under Cooperative Agreements NA17AE1625 and NA17AE1623, and by National Natural Science Foundation of China (NSFC) Grants 40605010, 40730950, and 41075041.

REFERENCES

  • Adler, R. F., and Coauthors, 2003: The version-2 Global Precipitation Climatology Project (GPCP) monthly precipitation analysis (1979–present). J. Hydrometeor., 4, 11471167.

    • Search Google Scholar
    • Export Citation
  • Awaka, J., , T. Iguchi, , and K. Okamoto, 1998: Early results on rain type classification by the Tropical Rainfall Measuring Mission (TRMM) precipitation radar. Proc. Eighth URSI Commission F Open Symp., Averior, Portugal, URSI, 134–146.

    • Search Google Scholar
    • Export Citation
  • Back, L. E., , and C. S. Bretherton, 2006: Geographic variability in the export of moist static energy and vertical motion profiles in the tropical Pacific. Geophys. Res. Lett., 33, L17810, doi:10.1029/2006GL026672.

    • Search Google Scholar
    • Export Citation
  • Berg, W., , C. Kummerow, , and C. Morales, 2002: Differences between east and west Pacific rainfall systems. J. Climate, 15, 36593672.

  • Cess, R. D., , M. Zhang, , B. A. Wielicki, , D. F. Young, , X. Zhou, , and Y. Nikitenko, 2001: The influence of the 1998 El Niño upon cloud radiative forcing over the Pacific warm pool. J. Climate, 14, 21292137.

    • Search Google Scholar
    • Export Citation
  • Chen, Y., , A. D. Del Genio, , and J. Chen, 2007: The tropical atmospheric El Niño signal in satellite precipitation data and a global climate model. J. Climate, 20, 35803601.

    • Search Google Scholar
    • Export Citation
  • Curtis, S., , and R. Adler, 2000: ENSO indices based on patterns of satellite-derived precipitation. J. Climate, 13, 27862793.

  • Dai, A., , and T. M. L. Wigley, 2000: Global patterns of ENSO-induced precipitation. Geophys. Res. Lett., 27, 12831286.

  • Del Genio, A. D., , and W. Kovari, 2002: Climatic properties of tropical precipitating convection under varying environmental conditions. J. Climate, 15, 25972615.

    • Search Google Scholar
    • Export Citation
  • DeMoss, J. D., , and K. P. Bowman, 2007: Changes in TRMM Rainfall due to the orbit boost estimated from buoy rain gauge data. J. Atmos. Oceanic Technol., 24, 15981607.

    • Search Google Scholar
    • Export Citation
  • Fu, Y., , and G. Liu, 2001: The variability of tropical precipitation profiles and its impact on microwave brightness temperatures as inferred from TRMM data. J. Appl. Meteor., 40, 21302143.

    • Search Google Scholar
    • Export Citation
  • Graham, N. E., 1995: Simulation of recent global temperature trends. Science, 267, 666671.

  • Houze, R. A., Jr., 1993: Cloud Dynamics. Academic Press, 573 pp.

  • Houze, R. A., Jr., 1997: Stratiform precipitation in regions of convection: A meteorological paradox? Bull. Amer. Meteor. Soc., 78, 21792196.

    • Search Google Scholar
    • Export Citation
  • Iguchi, T., , T. Kozu, , R. Meneghini, , J. Awaka, , and K. Okamoto, 2000: Rain-profiling algorithm for the TRMM Precipitation Radar. J. Appl. Meteor., 39, 20382052.

    • Search Google Scholar
    • Export Citation
  • Kubar, T. L., , and D. L. Hartmann, 2008: The vertical structure of tropical oceanic convective clouds and its relation to precipitation. Geophys. Res. Lett., 35, L03804, doi:10.1029/2007GL032811.

    • Search Google Scholar
    • Export Citation
  • Lambert, F. H., , S. Alexander, , Y. K. Nir, , and J. Chiang, 2008: How much will precipitation increase with global warming? Eos, Trans. Amer. Geophys. Union, 89, 21, doi:10.1029/2008EO210001.

    • Search Google Scholar
    • Export Citation
  • Li, R., , and Q.-L. Min, 2010: The impacts of mineral dust on the vertical structure of precipitation. J. Geophys. Res., 115, D09203, doi:10.1029/2009JD011925.

    • Search Google Scholar
    • Export Citation
  • Li, R., , Y. Fu, , and P. Zhao, 2005: Characteristics of rainfall structure over the tropical Pacific during the later period of 1997/1998 El Niño derived from TRMM PR observations (in Chinese). Chin. J. Atmos. Sci., 29, 225235.

    • Search Google Scholar
    • Export Citation
  • Lin, B., , B. A. Wielicki, , P. Minnis, , L. Chambers, , K. M. Xu, , Y. Hu, , and A. Fan, 2006: The effect of environmental conditions on tropical deep convective systems observed from the TRMM satellite. J. Climate, 19, 57455761.

    • Search Google Scholar
    • Export Citation
  • Liu, G., , and Y. Fu, 2001: The characteristics of tropical precipitation profiles as inferred from satellite radar measurements. J. Meteor. Soc. Japan, 79, 131143.

    • Search Google Scholar
    • Export Citation
  • Mapes, B. E., , and R. A. Houze Jr., 1995: Diabatic divergence profiles in western Pacific mesoscale convective systems. J. Atmos. Sci., 52, 18071828.

    • Search Google Scholar
    • Export Citation
  • Min, Q.-L., , R. Li, , B. Lin, , E. Joseph, , S. Wang, , Y. Hu, , V. Morris, , and F. Chang, 2009: Evidence of mineral dust altering cloud microphysics and precipitation. Atmos. Chem. Phys., 9, 32233231.

    • Search Google Scholar
    • Export Citation
  • Reynolds, R. W., , N. A. Rayner, , T. M. Smith, , D. C. Stokes, , and W. Wang, 2002: An improved in situ and satellite SST analysis for climate. J. Climate, 15, 16091625.

    • Search Google Scholar
    • Export Citation
  • Rogers, R. R., , and M. K. Yau, 1989: A Short Course in Cloud Physics. 3rd ed. Pergamon Press, 293 pp.

  • Ropelewski, C., , and M. Halpert, 1987: Global and regional scale precipitation patterns associated with the El Niño/Southern Oscillation. Mon. Wea. Rev., 115, 16061626.

    • Search Google Scholar
    • Export Citation
  • Satoh, S., , and A. Noda, 2001: Retrieval of latent heating profiles from TRMM radar data. Preprints, 30th Int. Conf. on Radar Meteorology, Munich, Germany, Amer. Meteor. Soc., 6.3. [Available online at http://ams.confex.com/ams/30radar/techprogram/paper_21763.htm.]

    • Search Google Scholar
    • Export Citation
  • Schumacher, C., , and R. A. Houze Jr., 2003: Stratiform rain in the tropics as seen by the TRMM precipitation radar. J. Climate, 16, 17391756.

    • Search Google Scholar
    • Export Citation
  • Schumacher, C., , R. A. Houze Jr., , and I. Kraucunas, 2004: The tropical dynamical response to latent heating estimates derived from the TRMM precipitation radar. J. Atmos. Sci., 61, 13411358.

    • Search Google Scholar
    • Export Citation
  • Shin, D.-B., , G. R. North, , and K. P. Bowman, 2000: A summary of reflectivity profiles from the first year of TRMM radar data. J. Climate, 13, 40724086.

    • Search Google Scholar
    • Export Citation
  • Short, D. A., , and K. Nakamura, 2000: TRMM radar observations of shallow precipitation over the tropical oceans. J. Climate, 13, 41074124.

    • Search Google Scholar
    • Export Citation
  • Soden, B. J., 2000: The sensitivity of the tropical hydrological cycle to ENSO. J. Climate, 13, 538549.

  • Solomon, S., , D. Qin, , M. Manning, , M. Marquis, , K. Averyt, , M. M. B. Tignor, , H. L. Miller Jr., , and Z. Chen, Eds., 2007: Climate Change 2007: The Physical Science Basis. Cambridge University Press, 996 pp.

    • Search Google Scholar
    • Export Citation
  • Steiner, M., , R. A. Houze Jr., , and S. E. Yuter, 1995: Climatological characterization of three-dimensional storm structure from operational radar and rain gauge data. J. Appl. Meteor., 34, 19782007.

    • Search Google Scholar
    • Export Citation
  • Tao, W. K., , J. Simpson, , S. Lang, , M. McCumber, , R. Adler, , and R. Penc, 1990: An algorithm to estimate the heating budget from vertical hydrometeor profiles. J. Appl. Meteor., 29, 12321244.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., 1998: Atmospheric moisture residence times and cycling: Implications for rainfall rates with climate change. Climatic Change, 39, 667694.

    • Search Google Scholar
    • Export Citation
  • Wang, P. H., , P. Minnis, , B. A. Wielicki, , T. Wong, , R. D. Cess, , M. H. Zhang, , L. B. Vann, , and G. S. Kent, 2003: Characteristics of the 1997/1998 El Niño cloud distributions from SAGE II observations. J. Geophys. Res., 108, 4009, doi:10.1029/2002JD002501.

    • Search Google Scholar
    • Export Citation
  • Wentz, F. J., , L. Ricciardulli, , K. Hilburn, , and C. Mears, 2007: How much more rain will global warming bring? Science, 317, 233235, doi:10.1126/science.1140746.

    • Search Google Scholar
    • Export Citation
  • Yang, S., , and E. A. Smith, 1999: Moisture budget analysis of TOGA COARE area using SSM/I-retrieved latent heating and large-scale Q2 estimates. J. Atmos. Oceanic Technol., 16, 633655.

    • Search Google Scholar
    • Export Citation
  • Yulaeva, E., , and J. M. Wallace, 1994: The signature of ENSO in global temperature and precipitation fields derived from the Microwave Sounding Unit. J. Climate, 7, 17191736.

    • Search Google Scholar
    • Export Citation
  • Zhang, C., 2001: Double ITCZs. J. Geophys. Res., 106 (D11), 11 78511 792.

Save