• Adler, R. F., , C. Kidd, , G. Petty, , M. Morrissey, , and H. M. Goodman, 2001: Intercomparison of global precipitation products: The third Precipitation Intercomparison Project (PIP-3). Bull. Amer. Meteor. Soc., 82 , 13771396.

    • Search Google Scholar
    • Export Citation
  • Arkin, P. A., , and B. N. Meisner, 1987: The relationship between large-scale convective rainfall and cold cloud over the Western Hemisphere during 1982–84. Mon. Wea. Rev., 115 , 297316.

    • Search Google Scholar
    • Export Citation
  • Arkin, P. A., , R. Joyce, , and J. E. Janowiak, 1994: The estimation of global monthly mean rainfall using infrared satellite data: The GOES precipitation index (GPI). Remote Sens. Rev., 11 , 107124.

    • Search Google Scholar
    • Export Citation
  • Berg, W., 2000: Differences in satellite estimates of precipitation over the east Pacific. Preprints, 10th Conf. on Satellite Meteorology and Oceanography, Long Beach, CA, Amer. Meteor. Soc., 375–377.

    • Search Google Scholar
    • Export Citation
  • Cess, R. D., , M. Zhang, , P. H. Wang, , and B. A. Wielicki, 2001: Cloud structure anomalies over the tropical Pacific during the 1997/98 El Niño. Geophys. Res. Lett., 28 , 45474550.

    • Search Google Scholar
    • Export Citation
  • Ferraro, R. R., , and G. F. Marks, 1995: The development of SSM/I rain-rate retrieval algorithms using ground-based radar measurements. J. Atmos. Oceanic Technol., 12 , 755770.

    • Search Google Scholar
    • Export Citation
  • Gruber, A., , X. Su, , M. Kanamitsu, , and J. Schemm, 2000: The comparison of two merged rain gauge–satellite precipitation datasets. Bull. Amer. Meteor. Soc., 81 , 26312644.

    • Search Google Scholar
    • Export Citation
  • Houghton, J. T., , L. G. Meira Filho, , B. A. Callender, , N. Harris, , A. Kattenberg, , and K. Maskell, Eds.,. . 1996: Climate Changes 1995: The Science of Climate Change. Cambridge University Press, 572 pp.

    • Search Google Scholar
    • Export Citation
  • Huffman, G. J., and Coauthors. 1997: The Global Precipitation Climatology Project (GPCP) combined precipitation dataset. Bull. Amer. Meteor. Soc., 78 , 520.

    • 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
  • Jaeger, L., 1983: Monthly and areal patterns of mean global precipitation. Variations in the Global Water Budget, A. Street-Perrott, M. Beran, and R. Ratcliffe, Eds., D. Reidel, 129–140.

    • Search Google Scholar
    • Export Citation
  • Janowiak, J. E., , P. A. Arkin, , P. Xie, , M. L. Morrissey, , and D. R. Legates, 1995: An examination of the east Pacific rainfall distribution. J. Climate, 8 , 28102823.

    • Search Google Scholar
    • Export Citation
  • Kummerow, C., 1993: On the accuracy of the Eddington approximation for radiative transfer in the microwave frequencies. J. Geophys. Res., 98 , 27572765.

    • Search Google Scholar
    • Export Citation
  • Kummerow, C., 1998: Beamfilling errors in passive microwave rainfall retrievals. J. Appl. Meteor., 37 , 356370.

  • Kummerow, C., , W. Barnes, , J. Shiue, , and J. Simpson, 1998: The Tropical Rainfall Measuring Mission (TRMM) sensor package. J. Atmos. Oceanic Technol., 15 , 809817.

    • Search Google Scholar
    • Export Citation
  • Kummerow, C., and Coauthors. 2001: The evolution of the Goddard profiling algorithm (GPROF) for rainfall estimation from passive microwave sensors. J. Appl. Meteor., 40 , 18011820.

    • Search Google Scholar
    • Export Citation
  • Legates, D. R., , and C. J. Wilmott, 1990: Mean seasonal and spatial variability in gauge-corrected, global precipitation. Int. J. Climatol., 10 , 111127.

    • Search Google Scholar
    • Export Citation
  • Petty, G. W., 1994a: Physical retrievals of over-ocean rain rate from multichannel microwave imagery. Part I: Theoretical characteristics of normalized polarization and scattering indices. Meteor. Atmos. Phys., 54 , 7999.

    • Search Google Scholar
    • Export Citation
  • Petty, G. W., 1994b: Physical retrievals of over-ocean rain rate from multichannel microwave imagery. Part II: Algorithm implementation. Meteor. Atmos. Phys., 54 , 101121.

    • Search Google Scholar
    • Export Citation
  • Petty, G. W., 1995: Frequencies and characteristics of global oceanic precipitation from shipboard present-weather reports. Bull. Amer. Meteor. Soc., 76 , 15931616.

    • Search Google Scholar
    • Export Citation
  • Petty, G. W., 1999: Prevalence of precipitation from warm-topped clouds over eastern Asia and the western Pacific. J. Climate, 12 , 220229.

    • 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.

  • Spencer, R. W., 1993: Global oceanic precipitation from the MSU during 1979–91 and comparisons to other climatologies. J. Climate, 6 , 13011326.

    • 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
  • Wentz, F. J., 1997: A well calibrated ocean algorithm for special sensor microwave/imager. J. Geophys. Res., 102 , 87038718.

  • Wentz, F. J., , and R. W. Spencer, 1998: SSM/I rain retrievals within a unified all-weather ocean algorithm. J. Atmos. Sci., 55 , 16131627.

    • Search Google Scholar
    • Export Citation
  • Wentz, F. J., , C. Gentemann, , D. Smith, , and D. Chelton, 2000: Satellite measurements of sea surface temperatures through clouds. Science, 288 , 847850.

    • Search Google Scholar
    • Export Citation
  • Wilheit, T. T., , A. T. C. Chang, , M. S. V. Rao, , E. B. Rodgers, , and J. S. Theon, 1977: A satellite technique for quantitatively mapping rainfall rates over the oceans. J. Appl. Meteor., 16 , 551560.

    • Search Google Scholar
    • Export Citation
  • Wilheit, T. T., , A. T. C. Chang, , and L. Chiu, 1991: Retrieval of monthly rainfall indices from microwave radiometric measurements using a probability distribution function. J. Atmos. Oceanic Technol., 8 , 118136.

    • Search Google Scholar
    • Export Citation
  • Xie, P. P., , and P. A. Arkin, 1997: Global precipitation: A 17-year monthly analysis based on gauge observations, satellite estimates, and numerical model outputs. Bull. Amer. Meteor. Soc., 78 , 25392558.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Mean DJF oceanic rainfall for the 10-yr period from 1987 to 1996 from available long-term satellite-based rain estimation techniques. The retrievals shown include (a) the GPI based on the technique developed by Arkin and Meisner (1987) and Arkin et al. (1994), (b) Spencer's (1993) estimates from the MSU, and (c) SSM/I emission-based retrievals from Wilheit et al. (1991). Boxes indicating the locations of the selected west (2°–12°N, 130°–160°E) and east Pacific (5°–15°N, 120°–150°W) regions are also shown

  • View in gallery

    Satellite-derived estimates of (a) SST from TMI (Wentz et al. 2000), (b) integrated water vapor from TMI (Wentz and Spencer 1998), (c) TMI rainfall (Kummerow et al. 2001), and (d) PR rainfall (Iguchi et al. 2000) for the 3-month period from DJF 1999/2000

  • View in gallery

    A comparison of mean profiles of (a) temperature, (b) temperature difference, and (c) relative humidity from ECMWF analysis for the selected east and west Pacific regions during DJF 1999/2000. Specified sea surface temperature values are from TMI (Wentz et al. 2000)

  • View in gallery

    Mean rainfall profiles from the TRMM PR for the selected east and west Pacific regions for DJF 1999/2000. Mean profiles are shown for both (a) stratiform and (b) convective rain systems

  • View in gallery

    Histograms of the relative contributions of convective, stratiform, and total rainfall from the TRMM PR for five different categories of surface rainfall rates during DJF of 1999/2000 over the selected east and west Pacific regions. Three pairs of bars corresponding to convective (C), stratiform (S), and total (T) rainfall are shown for each rain-rate category. For each pair the left-hand side bar (dark shading) corresponds to the west Pacific region while the right-hand side bar (light shading) is for the east Pacific. The far right-hand panel shows the totals for all five rain categories combined. Surface rain-rate values below 0.5 mm h−1 are excluded due to noise in the lightest rainfall cases

  • View in gallery

    Two-dimensional histograms of TMI scattering versus emission indices over the (a) west and (b) east Pacific regions. The histogram values include raining pixels only and they have been normalized by the total number of raining pixels in each region. The solid black line is the mean scattering index as a function of the emission index and the two dotted lines are provided for reference

  • View in gallery

    Simulated brightness temperatures as a function of cloud liquid water and rainfall computed using an Eddington radiative transfer model (Kummerow 1993) and a simple six-layer cloud in the Tropics. Results are shown for (a) the 50.3-GHz MSU channel used for the MSU precipitation retrievals (Spencer 1993) and (b) the 19.35-GHz channel, which is the primary emission channel used for many SSM/I and TMI retrievals. The dashed line shows the scattering effect of adding ice to the cloud above the freezing level

  • View in gallery

    Total rainfall (normalized) as a function of storm height over the selected east and west Pacific regions for DJF 1999/2000. The results are divided into separate categories for (a) small, (b) medium, and (c) large storm systems. Total rainfall within each 0.5-km storm height category has been normalized so that the area under the curve is the same for both regions and the maximum value is one

  • View in gallery

    A comparison of the mean DJF 1999/2000 height of the freezing level, or 0°C isotherm minus the height of the brightband or melting layer height from the TRMM PR. Results are shown for estimates of freezing level height from several sources including (a) NCEP reanalysis, (b) ECMWF analysis, and (c) retrievals from TRMM TMI using the technique by Wilheit et al. (1991)

  • View in gallery

    Mean rainfall profiles from the TRMM PR for the selected east and west Pacific regions for DJF 1997/98 during the intense El Niño event. Mean profiles are shown for both (a) stratiform and (b) convective rain systems

  • View in gallery

    Time series of rainfall anomalies from GPI and SSM/I (Wilheit et al. 1991). The mean rainfall values are averaged over the tropical oceans between 30°N and 30°S for (a) the west Pacific (110°E–160°W), (b) the east Pacific (160°–70°W), and (c) the global Tropics. For purposes of comparing the relative variability, the means have been removed and the values scaled by the standard deviation of the time series. A 13-month running mean filter has also been applied to the data

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 391 391 30
PDF Downloads 218 218 21

Differences between East and West Pacific Rainfall Systems

View More View Less
  • 1 Department of Atmospheric Science, Colorado State University, Fort Collins, Colorado
© Get Permissions
Full access

Abstract

A comparison of the structure of precipitation systems between selected east and west Pacific regions along the intertropical convergence zone (ITCZ) is made using a combination of satellite observations including vertical profile retrievals from the Tropical Rainfall Measuring Mission's (TRMM's) Precipitation Radar. The comparison focuses on the period from December 1999 to February 2000, which was chosen due to large discrepancies in satellite infrared and passive microwave rainfall retrievals. Storm systems over the east Pacific exhibit a number of significant differences from those over the west Pacific warm pool including shallower clouds with warmer cloud tops, a larger proportion of stratiform rain, less ice for similar amounts of rainwater, and a radar bright band or melting layer significantly farther below the freezing level.

These regional differences in the structure of precipitation systems between the east and west Pacific also exhibit seasonal and interannual variability. During the intense 1997/98 El Niño, warmer sea surface temperatures (SSTs) in the east Pacific led to precipitation systems with a very similar structure to those observed over the west. These differences in east versus west Pacific rainfall and changes associated with the El Niño–Southern Oscillation (ENSO) result in time-dependent regional biases in available long-term satellite precipitation datasets. Although all of the currently available infrared and passive microwave–based satellite retrievals exhibit similar spatial patterns and capture variability associated with ENSO, both the amplitude and sign of subtle climate signals, such as the response of tropical-mean rainfall to ENSO, depend on the retrieval algorithm used.

Corresponding author address: Wesley Berg, Dept. of Atmospheric Science, Colorado State University, Fort Collins, CO 80523-1371. Email: berg@atmos.colostate.edu

Abstract

A comparison of the structure of precipitation systems between selected east and west Pacific regions along the intertropical convergence zone (ITCZ) is made using a combination of satellite observations including vertical profile retrievals from the Tropical Rainfall Measuring Mission's (TRMM's) Precipitation Radar. The comparison focuses on the period from December 1999 to February 2000, which was chosen due to large discrepancies in satellite infrared and passive microwave rainfall retrievals. Storm systems over the east Pacific exhibit a number of significant differences from those over the west Pacific warm pool including shallower clouds with warmer cloud tops, a larger proportion of stratiform rain, less ice for similar amounts of rainwater, and a radar bright band or melting layer significantly farther below the freezing level.

These regional differences in the structure of precipitation systems between the east and west Pacific also exhibit seasonal and interannual variability. During the intense 1997/98 El Niño, warmer sea surface temperatures (SSTs) in the east Pacific led to precipitation systems with a very similar structure to those observed over the west. These differences in east versus west Pacific rainfall and changes associated with the El Niño–Southern Oscillation (ENSO) result in time-dependent regional biases in available long-term satellite precipitation datasets. Although all of the currently available infrared and passive microwave–based satellite retrievals exhibit similar spatial patterns and capture variability associated with ENSO, both the amplitude and sign of subtle climate signals, such as the response of tropical-mean rainfall to ENSO, depend on the retrieval algorithm used.

Corresponding author address: Wesley Berg, Dept. of Atmospheric Science, Colorado State University, Fort Collins, CO 80523-1371. Email: berg@atmos.colostate.edu

1. Introduction

It has been suggested that in future climate scenarios warmer temperatures associated with increases in greenhouse gases will lead to a more vigorous hydrological cycle and possibly an increase in precipitation intensity resulting in more extreme rainfall events (Houghton et al. 1996). To attempt to evaluate the validity of this statement, as part of the Atmospheric Modeling Intercomparison Project (AMIP), Soden (2000) performed comparisons between a set of 30 general circulation models and observations based on satellite retrievals for the period from 1979 to 1988. Although this period is too short to evaluate potential changes associated with any climate trends, Soden suggested that climate variability associated with the El Niño–Southern Oscillation (ENSO) serves as a useful test bed for assessing coupling between the components of the hydrologic cycle and thus evaluating climate model performance. As a result of these comparisons, he found that the model-predicted changes in tropical-mean precipitation associated with ENSO were a factor of 4 less than the observed values, although both the models and observations showed an increase in precipitation during El Niño. This result led him to suggest that either existing climate models are substantially under predicting changes in the hydrological cycle resulting from increases in sea surface temperature (SST) associated with ENSO, or that satellite observations of tropical-mean (oceanic) precipitation are inadequate to the task of monitoring ENSO-related changes in precipitation.

The precipitation “observations” used in Soden's (2000) comparison were actually rainfall estimates produced by Spencer (1993) using data from the microwave sounding unit (MSU). Janowiak et al. (1995) performed a comparison of climatological precipitation datasets from in situ and satellite sources including the MSU retrievals from Spencer (1993). They found substantial disagreement in seasonal rain rates for the period December–January–February (DJF) over the extreme east Pacific intertropical convergence zone (ITCZ) (90°–100°W). Mean values ranged from 2 mm day−1 for infrared-based Geostationary Operational Environmental Satellite (GOES) precipitation index (GPI) estimates (Arkin and Meisner 1987) to 10 mm day−1 for the Spencer MSU estimates. Looking at circulation data from the U.S. National Meteorological Center [(NMC), now known as the National Centers for Environmental Prediction (NCEP)] analysis they found relatively shallow divergence in this region leading them to infer the presence of shallow convection capped near the 500-mb level. Janowiak et al. (1995) also compared these same datasets over the east-central Pacific (120°–160°W). Here they found that passive microwave rainfall estimates show significantly more rainfall than that indicated by either the GPI estimates or NMC circulation data. They concluded from this that as much as 50%–75% of the rainfall in this region is generated from relatively shallow cloud systems. A more recent analysis of tropical oceanic rain systems using the Tropical Rainfall Measuring Mission (TRMM) Precipitation Radar (PR) by Short and Nakamura (2000) found evidence of bimodel storm height distributions with a mode of shallow precipitation around 2–3 km in height contributing around 21% of the total rainfall during El Niño and 22% during La Niña conditions.

Petty (1995) investigated regional variations in oceanic rainfall using shipboard present-weather reports. Based on his comparisons he found a strong regional dependence in the frequency of thunderstorms as well as other rain types leading him to conclude that “it seems unlikely that any satellite technique can be made to yield unbiased results globally without the benefit of careful region-by-region calibration.” Petty (1999) also investigated the prevalence of rainfall by warm-topped clouds using coincident satellite infrared images from the Japanese Geostationary Meteorological Satellite (GMS-5) with land- and ship-based synoptic reports of precipitation. Over the South China Sea and the surrounding region he found evidence of between 10% and 50% of precipitation reports associated with minimum IR cloud-top temperatures above 273 K. He also found evidence of significant seasonal variability in the fraction of precipitation from warm-topped clouds leading him to suggest “the possibility of significant regional and seasonal biases in satellite rainfall estimates based on algorithms and sensors sensitive primarily to cold-topped clouds.” Similar comparisons of geostationary infrared imagery from both GMS-5 and GOES with precipitation estimates from the Special Sensor Microwave Imager (SSM/I; Berg 2000) also showed a significant contribution to tropical precipitation by warm-topped clouds over the east Pacific ITCZ. Large seasonally varying differences were found in the amount of warm precipitation between the east and west Pacific with substantially more warm rain in the east during DJF as suggested by Janowiak et al. (1995). This evidence of potentially large seasonally varying regional biases in infrared satellite rainfall retrievals raises serious questions regarding the validity of the observed variability resulting from changes associated with ENSO in rainfall datasets utilizing IR retrievals, at least with those techniques using fixed thresholds such as the GPI. It also raises questions regarding the possibility of biases in MSU and other passive microwave rainfall estimates.

Using recent data from TRMM, including observations of the vertical structure of rain systems observed with the PR, we attempt to investigate differences between east and west Pacific rain systems as well as changes associated with the intense 1997/98 El Niño event. The TRMM satellite was launched in November of 1997 for the purpose of measuring rainfall and energy exchange over tropical and subtropical regions. The satellite is in a 350-km circular orbit with an inclination of 35°. The three rainfall sensors on board include the TRMM Microwave Imager (TMI), the PR, and the Visible and Infrared Scanner (VIRS). These sensors provide information on the three-dimensional structure of precipitation and heating in the Tropics, including rainfall vertical structure information from the TRMM PR. The PR operates at a frequency of 13.8 GHz scanning a 215-km swath with a horizontal resolution of 4.3 km at nadir and a vertical resolution of 250 m. Additional information on the TRMM satellite including specifications for the sensors on board is provided by Kummerow et al. (1998).

2. Satellite climate rainfall datasets

A number of large-scale precipitation datasets covering a decade or more are currently available. Although satellite-based retrievals offer the only source of large-scale precipitation observations over most ocean regions, two merged datasets combining a variety of satellite rainfall estimates with rain gauge measurements have been produced. These include the Global Precipitation Climatology Project (GPCP; Huffman et al. 1997) and the Climate Prediction Center's (CPC) Merged Analysis of Precipitation (CMAP; Xie and Arkin 1997). While the GPCP dataset was originally produced for the period of available SSM/I data (from mid-1987 forward), it has been extended using additional datasets to encompass the period from January 1979 to the present. The CMAP data covers the same period, although the component satellite datasets used in both the merged products vary with time due to availability. Both the merged rainfall datasets incorporate measurements from rain gauges with satellite estimates over land, while over ocean regions they rely on a combination of satellite infrared and passive microwave rainfall estimates. Infrared-based retrievals are included from both geostationary and polar-orbiting satellites using the GPI technique (Arkin and Meisner 1987; Arkin et al. 1994). Passive microwave retrievals from the SSM/I include both emission (Wilheit et al. 1991) and scattering-based (Ferraro and Marks 1995) rainfall estimates. The GPCP dataset adjusts the IR rainfall estimates based on passive microwave retrievals while the CMAP dataset weighs the individual datasets according to their correlations with gauge data. In addition, CMAP uses the Spencer (1993) MSU estimates for the period prior to the launch of SSM/I in 1987.

An intercomparison of global precipitation products by Adler et al. (2001) investigated differences between a number of climate rainfall products including 25 observational (satellite based), 4 model, and 2 climatological datasets. They concluded that the results of the intercomparison established the value of the merged products for applications requiring global monthly precipitation. They also found that interannual results from the observational products were highly variable and, therefore, recommended that the data should be used with caution when assessing changes associated with interannual variability. A comparison of the two merged datasets, CMAP and GPCP, by Gruber et al. (2000) for the original period of the SSM/I-era GPCP data (July 1987–December 1998) found that the spatial and temporal variability captured by the satellite retrievals shows generally good agreement and is a substantial improvement over mean climatologies based on gauge and ship measurements (Jaeger 1983; Legates and Wilmott 1990) or general circulation model (GCM) analyses. They found that the two datasets were highly correlated both in terms of the resulting spatial patterns and their depiction of patterns of interannual variability associated with ENSO. They did, however, find differences in the mean values, the global distribution, and zonal average time series. Although these differences are more subtle, they become important when looking for relatively small climate signals.

To address the question raised by Soden (2000) regarding the utility of satellite rainfall estimates for monitoring relatively small climate signals, however, a better understanding of the sensitivity of these individual satellite retrievals to regional and/or temporal changes in precipitation systems is needed. Because this is a complicated and difficult task, we have chosen to focus on the previously observed differences between east and west Pacific rainfall systems (Janowiak et al. 1995), which also exhibit substantial variability associated with ENSO. The questions we wish to address are how do differences between east and west Pacific rainfall affect the long-term infrared and passive microwave satellite retrievals described above and can we ultimately use the satellite-based observations to monitor changes in tropical-mean rainfall associated with ENSO?

3. East versus west Pacific rainfall

Figure 1 shows a comparison of mean rainfall from infrared and passive microwave satellite data during the Northern Hemisphere winter season (DJF) averaged over a 10-yr period from 1987 to 1996. With the exception of the MSU estimates, which are not incorporated into the GPCP product, these datasets are the primary component satellite estimates used over tropical oceans in the long-term GPCP and CMAP rainfall datasets. The Northern Hemisphere winter was chosen for this comparison because, during DJF, differences between infrared and passive microwave retrievals are at a maximum over the east Pacific. The figure clearly shows large differences in east Pacific ITCZ rainfall, with the MSU indicating 3–4 times the amount of rainfall as the IR-based GPI estimate. To determine the underlying reasons for these large discrepancies we have examined a variety of observed quantities related to the structure of storm systems beginning with differences in the background states between the east and west Pacific.

a. Differences in the background states of the east and west Pacific

Two regions along the ITCZ were selected for comparison focusing on a single DJF season. One region was selected over the west Pacific warm pool region and the other encompassing the region of maximum differences between infrared and passive microwave rainfall retrievals in the central/eastern Pacific as shown in Fig. 1. Both of the selected regions are of the same dimensions and are over ocean-only areas along the ITCZ. Boxes drawn around the selected regions are also shown in Fig. 1. The season selected for the comparison was from December 1999 to February 2000 (DJF 1999/2000) because of the relatively “normal” SSTs during this period. Although weak to moderate La Niña conditions existed at this time, the resulting rainfall patterns appeared very similar to normal conditions. This is in stark contrast to the December 1997–February 1998 period, which was during the intense 1997/98 El Niño. Figure 2 shows a comparison over the Pacific of mean sea surface temperature, total column water vapor, and rainfall from both the TMI and the PR for DJF 1999/2000. Enlarged images of the two regions are shown on the right-hand side of Fig. 2.

Sea surface temperature estimates from TMI (Wentz et al. 2000), shown in Fig. 2, indicate an average temperature of 25.2°C for the selected east Pacific region versus 29.3°C over the west Pacific, a difference of 4.1°. Substantially more water vapor is associated with the warmer SSTs over the west and the rainfall patterns indicate more widespread precipitation over the warm pool with a narrower band of precipitation in the east. The TMI and PR estimates exhibit differences in small-scale variability due to sampling and other issues; however, the overall patterns appear very similar. While a positive bias of around 30% over the Tropics for TMI estimates relative to the PR has previously been documented for the version 5 products shown here (Kummerow et al. 2001), this bias does appear to be slightly larger for the selected east Pacific region than over the warm pool. Mean rainfall averaged over the specified regions are 7.09 over the west versus 4.73 mm day−1 over the east from the TMI retrieval and 5.57 over the west versus 3.50 mm day−1 over the east from the PR retrieval. The larger mean rainfall values over the west Pacific shown in Fig. 2 are primarily due to the areal coverage by rainfall events. In the east Pacific region precipitation occurs mainly over a narrow band along the ITCZ leaving larger areas within the region with zero or minimal rain.

Figure 3 shows mean atmospheric temperature and relative humidity profiles for the two regions based on analysis from the European Centre for Medium-Range Weather Forecasts (ECMWF). As expected due to the colder SSTs in the east Pacific, the temperature profile is significantly colder throughout the lower troposphere. The temperature difference between the two regions is shown in Fig. 3b. This figure indicates that the large temperature difference near the surface decreases with height up to around 3 km due to a lower lapse rate in the east Pacific.

According to the ECMWF analysis the height of the freezing level is over 500 m higher in the west Pacific than in the east. This is around 200 m less than expected based on the difference in SST alone and assuming a constant lapse rate of 6°C km−1, but is a result of the lower lapse rate in the east as shown in Fig. 3b. Figure 3c shows the relative humidity profiles from ECMWF for the two regions, indicating drier air aloft around and above the freezing level in the east Pacific. While the ECMWF analysis is useful to look at differences between the mean vertical temperature and moisture structures for the two regions, the analysis lacks in situ observations over the east Pacific, although satellite temperature and moisture information is assimilated into the model during this period. Comparing the ECMWF analysis shown here with corresponding analysis from the National Centers for Environmental Prediction (NCEP; not shown) reveals significant differences in the temperature and moisture profiles, most notably with regard to the humidity at upper levels. According to the NCEP analysis, however, the height of the freezing level is only 350 m higher in the west than in the east, indicating an even larger discrepancy in the east versus west Pacific lapse rates than indicated by the ECMWF analysis. While the EMCWF analysis assimilates satellite observations not used in the NCEP analysis, it is clear that these results should be viewed with some skepticism. Both the ECMWF and NCEP analysis, however, support a presumption of more low-level heating associated with an increase in shallow rain systems as well as large-scale subsidence and drier air aloft over the east Pacific region.

b. Comparison of mean rainfall profiles

Mean rainfall profiles from the TRMM PR averaged over DJF 1999/2000 are shown for the selected east and west Pacific regions in Fig. 4. Separate profiles are shown for stratiform and convective rainfall as determined by the classification technique developed by Steiner et al. (1995), which is used in the PR retrieval algorithm (Iguchi et al. 2000). The mean stratiform profiles, in Fig. 4a, indicate deeper clouds with a stronger bright band in the west. Although the PR profile indicates more water aloft in the west Pacific, the surface rain rates are very similar with a mean of 1.93 in the west and 1.97 mm h−1 in the east. The convective profiles also indicate higher clouds in the west Pacific with more liquid water at upper levels; however, the mean surface rain rate is larger in the east at 8.61 versus 7.25 mm h−1 for the west. This suggests that although the mean rain profile is shallower in the east the mean surface rain rate is actually larger. This result also supports the results of Janowiak et al. (1995) and Berg (2000), which found a larger contribution by shallow systems to the total rainfall over the east Pacific. The stratiform/convective breakdown in the east Pacific is 54% stratiform to 46% convective, which is almost opposite the breakdown in the west. As a result, there is about 10% more stratiform rainfall in the east than the west, although it is possible that this is partly due to the classification scheme used in the PR retrieval algorithm related to shallow clouds not reaching the freezing level.

A histogram of the relative contribution of stratiform, convective, and total rainfall systems divided into categories based on the surface rain rate is shown for the selected east and west Pacific regions in Fig. 5. Comparing the total rainfall, or convective and stratiform combined, reveals more low intensity (0.5–2.0 mm h−1) and high intensity (>20 mm h−1) systems in the east Pacific. While one might expect that shallower stratiform systems in the east would lead to more light rainfall, the contribution by intense rainfall systems is larger in the east as well. The presence of more stratiform and less convective rain over the east Pacific is evident through all the rain-rate categories with the exception of the most intense systems where the east Pacific has a larger contribution by convective systems. As expected, convective systems dominate the intense rainfall category and most of the light rain cases are classified as stratiform.

c. Ratio of liquid to ice content

Utilizing data from the TRMM TMI, a comparison of the 19-GHz emission or attenuation index versus the 85-GHz scattering index was made for the two regions based on the method of Petty (1994a). The emission index measures the difference between the horizontally and vertically polarized brightness temperatures at 19 GHz relative to that of the highly polarized ocean surface background, which is related to the attenuation by liquid hydrometeors and, thus, to the total columnar liquid water content. The 85-GHz scattering index measures the amount of radiation scattered out of the satellite field of view, which is primarily due to frozen precipitation aloft at 85 GHz and thus to the total columnar ice content. Figure 6 shows two-dimensional histogram plots of the scattering versus emission indices for raining TMI pixels over the west and east Pacific regions. For purposes of comparison the values have been normalized and a nonlinear grayscale applied. While the scattering index increases with increasing rainfall rate, the emission index decreases with increasing hydrometeor content. For this reason, the scattering index is plotted as a function of 1 − emission index. The mean scattering index for a given emission index is plotted with a solid black line and dotted reference lines are included to aid in comparing differences between the two regions. For the west Pacific region, the overall shape of the histogram and the increased frequency of occurrence of larger scattering indices, shown between the dotted lines, indicates that for a given emission index, or liquid water content, there is significantly more scattering and thus more ice in the column. It should be noted that a systematic regional change in the beam-filling effect resulting from inhomogeneous rainfall could also be partly responsible for the differences shown in Fig. 6. This is due to the fact that the spatial resolution of the 19-GHz channels used to compute the emission index is substantially larger than that of the 85-GHz channels used to compute the scattering index. Preliminary comparisons indicate that while this effect is not insignificant, it is relatively small. The presence of more ice in the column, however, is consistent with the deeper cloud profiles shown in Fig. 4. Given less ice in the atmospheric column for the same liquid water emission signal, passive microwave scattering–based algorithms will tend to underestimate rainfall in the east Pacific relative to algorithms relying on an emission or attenuation signal by liquid hydrometeors.

This decrease in the total ice content relative to the liquid water content for precipitating storm systems over the east Pacific may also have a significant effect on the MSU rainfall estimates shown in Fig. 1. Suggested by F. R. Robertson, E. W. McCaul, and D. E. Fitzjarrald (1999, personal communication), an overestimation by the MSU technique is a result of the effects of scattering by ice particles being neglected in the original algorithm formulation. Figure 7 shows a comparison of the effect of ice on simulated brightness temperatures for the 50.3-GHz MSU channel 1 used in the MSU retrieval algorithm and the 19.35-GHz channel used by various SSM/I and TMI emission-based retrieval algorithms. The results shown in the figure are based on radiative transfer calculations from an Eddington model (Kummerow 1993) using a simple six-layer cloud with homogeneous rain/ice content within the field of view. The simulated cloud contains a constant liquid water content corresponding to the specified rain rate from the surface up to 4.5 km then decreasing to half that value at the top of the 9-km cloud. For the case with ice, a mixed layer of ice and water is included from 4.5 to 6 km with an all-ice phase between 6 and 9 km. While the simulated cloud is rather simplistic, it is clear that the MSU channel is significantly more sensitive to ice than the lower-frequency SSM/I channel. Spencer (1993) acknowledges the potential brightness temperature depression due to precipitation-sized ice particles, but concludes that the low 100–200-km spatial resolution of the MSU instrument makes the effect of scattering due to ice relatively insignificant. As Fig. 7 shows, however, there is a significant depression in the 50.3-GHz brightness temperatures for moderate rain rates. This suggests that intense convective storms only partially filling the MSU footprint would still have a significant impact on the observed brightness temperatures.

As Fig. 7 shows, the warming effect of 0.5 kg m−2 cloud water in the column produces an increase of 19° in the 50.3-GHz MSU channel versus a 15.8° increase in the 19.35-GHz channel. More importantly, the brightness temperature increase associated with cloud water accounts for approximately 50% of the maximum warming in the simulated rain cloud for the MSU channel, while it only accounts for around 14% of the maximum warming for the 19.35-GHz channel. This means that changes in cloud water content will have a substantially larger effect on the MSU retrieval algorithm than on an emission-based retrieval utilizing the 19.35-GHz channel. While Spencer (1993) recognized that the warming of the brightness temperatures due to cloud water may be significant, he addressed this issue by calibrating the MSU emission signal using measurements from a sparse collection of rain gauges distributed around the globe. This global calibration, however, will not account for regional differences in cloud water content. Comparisons of cloud liquid water retrievals from SSM/I (Wentz 1997) between the selected east and west Pacific regions do not indicate a systematic difference; however, it is a difficult quantity to accurately retrieve and almost impossible to distinguish from low-intensity rainfall. The presence of more light rain events over the east, as shown in Fig. 5, may be related to an overall increase in cloud liquid water, but based on currently available data this effect would be extremely difficult to quantify.

While the calibration of the MSU emission signal, based on global rain gauge data, will presumably remove a mean global systematic bias due to the ice scattering effect, any regional east Pacific bias will remain. Since no gauge data is available over the east Pacific, the calibration will adjust the results to the mean conditions of the gauge locations, which are primarily in the west Pacific and coastal regions (Spencer 1993). Because the east Pacific has significantly less ice for comparable rainwater content than in the west Pacific, where many of the gauges used for the calibration are located, a smaller depression in the brightness temperatures will occur resulting in artificially inflated rainfall estimates. Figure 1 shows a high bias in east Pacific MSU rainfall relative to the SSM/I emission-based retrievals, which would appear to support this hypothesis. Of course, other factors such as regional differences in the spatial inhomogeneity of rainfall (Kummerow 1998) or cloud water content may also contribute to this difference. The fact that the primary emission channel used in the SSM/I retrieval (Wilheit et al. 1991) shows almost no sensitivity to ice, however, suggests that the difference in ice content between the east and west may be the source of the apparent high bias in MSU rainfall estimates over the east Pacific.

d. Distribution of storm system types

As mentioned previously, Janowiak et al. (1995) and Berg (2000) found evidence that the east Pacific has shallower precipitation systems with warmer cloud tops, thus, leading to an underestimate by infrared retrieval techniques. Using infrared imagery from the TRMM VIRS in conjunction with TRMM PR profile retrievals, variations in the height of rain systems were investigated as a function of storm system size. Histograms of the height of the storm top obtained from the PR 2A25 dataset as a function of storm system size are shown in Fig. 8. A simple classification into small, medium, and large storm systems was made based on the number of contiguous infrared pixels colder than 280 K. Small storm systems were defined as those with an areal coverage by contiguous infrared pixels below 280 K of less than 1600 km2, medium systems were identified as those covering an area between 1600 and 16 000 km2, and large systems were specified as covering an area over 16 000 km2. The histograms shown in Fig. 8 indicate a shift toward shallower systems in the east Pacific for all sizes of storm systems. This coincides with the mean profile differences shown in Fig. 4. It is significant, however, that the biggest difference in the distribution of storm height between the two regions exists for small storm systems. Over the east Pacific, small systems produce 7% of the total rain and have a median rain column height of around 4 km versus in the west where small systems produce 10% of the total and have a median rain column height between 6.5 and 7 km. Medium size storm systems are responsible for 35% of the total rain in the east versus 31% for the west, and the large systems for 58% in the east versus 59% in the west.

An important result demonstrated by the storm height distributions shown in Fig. 8 is that there is substantial overlap between the two regions. This means that tall storm systems exist in the east Pacific and shallow systems occur in the west. The most significant differences are in the mean storm heights and the resulting shape of the distributions. As a result, a limited set of observations from a field program may not accurately reflect a shift in the distribution. Since IR cloud-top temperature is closely related to storm height, it is apparent from Fig. 8 that a technique like the GPI retrieval, which relates the fraction of clouds below 235 K to the total rain, will significantly underestimate rain in the east due to an increased frequency of shallow precipitation systems.

e. Freezing height versus PR brightband height

Emission-based passive microwave retrieval algorithms relate brightness temperature increases resulting from the emission by liquid hydrometeors to the rainfall rate (e.g., Wilheit et al. 1991; Petty 1994a). Because the frequencies used are relatively transparent to the clear-sky atmosphere, the emission signal provides an excellent measure of the total column liquid water content. To compute the surface rain rate, therefore, it is necessary to assume or independently determine both the shape of the liquid water profile and the height of the liquid water column. Petty (1994b) computes an estimate of the freezing level height from the brightness temperatures based on a relationship between column water vapor and freezing level. Two of the standard TRMM TMI ocean retrieval algorithms, the 2A12 Goddard profiling algorithm (GPROF; Kummerow et al. 2001) and the 3A11 algorithm (Wilheit et al. 1991), also calculate the height of the liquid water column directly from the observed brightness temperatures. The 2A12 and 3A11 algorithms estimate the height of the freezing level using a method developed by Wilheit et al. (1991), which relates the 19- and 22-GHz vertically polarized brightness temperatures to the freezing height based on radiative transfer model calculations (Wilheit et al. 1977). This method relies on the relationship between the total column water vapor content for saturated (i.e., raining) conditions and the height of the freezing level. Although there are potential problems associated with this approach, it does appear to capture the lower freezing level heights associated with colder SSTs in the east Pacific. Other emission algorithms use climatological freezing heights or a fixed height (e.g., Wentz and Spencer 1998) to determine the rain layer thickness. Although all emission algorithms require an estimate of the rain layer thickness, some algorithms may do this implicitly through empirically derived algorithm coefficients.

Because of limited radiosonde coverage over regions such as the east Pacific it is difficult to obtain an accurate estimate of the height of the freezing level. For stratiform rainfall, however, a more accurate measure of the height of the liquid water column is the height of the PR brightband. The brightband height provides an actual measure of the mean location where ice particles are melting in stratiform rain systems. As a result, it provides a better estimate of the rain layer thickness. The difference between the height of the freezing level, or zero degree isotherm, with the observed brightband height from the TRMM PR is shown in Fig. 9. Results are shown for freezing height estimates from three different sources, including NCEP reanalysis, ECMWF analysis, and retrieved values from TMI using the Wilheit et al. (1991) technique. Because the PR brightband height is used for Figs. 9a–c, differences between the three are due to differences in the estimated height of the freezing level between NCEP, ECMWF, and TMI. What is most interesting, however, is the systematic difference between the east and west Pacific regardless of the freezing height estimate that is used.

Comparing the results for the TMI-derived freezing level height with the PR brightband height shown in Fig. 9c, the mean distance of the brightband below the freezing level is around 700 m in the east Pacific versus only 500 m in the west. The discrepancy between the east and west using the NCEP-derived freezing level height, shown in Fig. 9a, is even larger. The reason for this large difference between the two regions is unclear. If this difference is real, and not simply an artifact from the models, it may be related to the Walker circulation, which produces large-scale subsidence and thus drier air over the east Pacific. Simulations of melting particles in stratiform rain systems suggest that dry air near the freezing level is the only realistic way of producing a brightband that is as much as 800 m below the freezing level (Battaglia et al. 2000, manuscript submitted to J. Appl. Meteor.). Although a storm system will result in locally saturated conditions, mean clear-sky conditions with drier air in the east Pacific, as shown in Fig. 3, may influence this difference between the height of the freezing level and the brightband. Unfortunately, a lack of radiosonde observations within stratiform rain systems makes it very difficult to find observational evidence either confirming or refuting this hypothesis. Regardless of the source, Fig. 9c indicates a significant regional difference in the observed (PR brightband height) versus the estimated (TMI-retrieved freezing height) values for the height of the liquid water column between the east and west Pacific. This will result in a relative bias in surface rainfall between the east and west Pacific by emission-based passive microwave retrievals, such as SSM/I and TMI, which utilize the Wilheit et al. (1991) freezing height estimates as an indicator of the liquid water column height.

4. The effects of El Niño and La Niña

Using cloud data from the Stratospheric Aerosol and Gas Experiment (SAGE) II, Cess et al. (2001) found a substantial change in cloud vertical structure over the tropical Pacific during the intense 1997/98 El Niño. They found that relative to normal years, clouds were lower in the west Pacific and higher in the east, which they attributed to a collapse of the Walker circulation. A comparison of east versus west Pacific storm structures during DJF 1997/98 using the TRMM PR retrievals is shown in Fig. 10 for both stratiform and convective rain systems. During this period the SSTs were very similar over the selected east and west Pacific regions with a mean temperature of 29.4°C in the west and 29.1°C in the east. Compared to the profiles shown in Fig. 4 for DJF 1999/2000, the mean rain profiles, freezing heights, and brightband heights are far more similar between the two regions during the El Niño. The most significant difference shown in Fig. 10 is slightly more intense stratiform systems over the east Pacific. The large discrepancy in the height of the rain layer evident during DJF 1999/2000 has decreased significantly with only slightly deeper systems occurring over the east Pacific. Both regions show an increase in the percentage of stratiform systems relative to the normal conditions during DJF 1999/2000 with the east Pacific continuing to have a higher occurrence of stratiform rain. Many of the other characteristics distinguishing the structure of east Pacific rainfall systems from the west are also greatly reduced during El Niño. As a result, many of the systematic east–west Pacific satellite retrieval biases discussed previously decrease or even disappear during the El Niño. Because these regionally dependent systematic biases in the GPI, MSU, and even SSM/I retrievals vary with the occurrence of El Niño, the interannual variability of tropical-mean rainfall as measured by these various techniques is also affected.

Figure 11 shows a comparison of the time series of tropical-mean oceanic precipitation from infrared-based GPI estimates (Arkin and Meisner 1987) and passive microwave emission–based SSM/I estimates (Wilheit et al. 1991). The time series are shown for both the east and west Pacific as well as the global tropical oceans. A much broader definition for the east and west Pacific regions was used for this comparison. The intense 1997/98 El Niño is clearly present in the east Pacific time series of both datasets. Warmer SSTs in the east during this event led to a significant increase in precipitation over the region, although the amplitude of this increase is larger in the SSM/I estimates. A corresponding decrease in precipitation occurs over the west Pacific during El Niño, which is also evident in both datasets. It is the relative magnitude of the increase (decrease) in rainfall between the east (west) Pacific, however, that determines the sign and amplitude of the global precipitation response to El Niño. The GPI dataset shows a decrease in tropical-mean precipitation associated with El Niño while the SSM/I dataset shows a marked increase. It is apparent from Fig. 11 that the GPI and SSM/I estimates, which exhibit large differences in the east–west gradient of precipitation, as shown in Fig. 1, both detect the increase (decrease) in precipitation over the east (west) Pacific associated with El Niño. The impact of the 1997/98 El Niño on the combined tropical-mean signal, however, differs in both amplitude and sign between the two datasets. While this may be due to limitations of these specific retrieval algorithms, a similar comparison of a 4-yr time series from the current TRMM PR and TMI retrieval algorithms (not shown here) also indicates a significant difference in the response of tropical-mean rainfall to the 1997/98 El Niño. Although corresponding MSU precipitation estimates are not available during this event, the interannual variability in east versus west Pacific biases affecting the GPI, MSU, and SSM/I retrievals raise serious doubt as to the validity of the large increase in tropical-mean MSU-derived precipitation discussed in Soden's (2000) comparison. Simply put, the various satellite precipitation retrievals examined here are not currently of sufficient accuracy to detect the relatively small climate signals associated with variability in tropical-mean precipitation.

5. Discussion and conclusions

During DJF 1999/2000 in the absence of a significant El Niño event, precipitation systems in the east Pacific were shallower than those in the west Pacific, especially for small storm systems, and had a higher percentage of stratiform rain. West Pacific storm systems contained more ice in the atmospheric column for a comparable amount of liquid water, while in the east colder SSTs contributed to a lower freezing height and shallower rain systems. Systematically lower brightband heights in stratiform systems over the east Pacific relative to various estimates of the freezing level height suggest dynamical influences on the mean rainfall profile over the east. During DJF 1997/98, which corresponded to an intense El Niño event, the mean structure of rain systems appeared more similar between the selected east and west Pacific regions; however, differences such as a larger fraction of stratiform rain systems over the east Pacific remained. There was also a significant increase in the fraction of stratiform rain relative to DJF 1999/2000 over both the east and west Pacific regions, even though the mean SST over the selected west Pacific region was almost identical to that observed during the DJF 1999/2000 period. While the changes in SST associated with the 1997/98 El Niño clearly had a significant impact on the structure of Pacific rainfall systems, corresponding changes in the large-scale circulation patterns also appear to be important. Specifically, the Walker circulation may play a significant role in modifying the mean structure of rain systems along the east Pacific ITCZ.

These distinct differences in the structure of precipitation systems between the east and west Pacific and changes in the structure associated with El Niño lead to time-dependent regional biases in the various satellite rainfall estimates. Shallower rain systems in the east result in an underestimate of rainfall by infrared retrieval techniques, which relate the occurrence of high cold clouds to rainfall. The presence of less ice in east Pacific storms for comparable emission signatures, which is related to the liquid water content and thus more directly to the surface rainfall rate, will potentially lead to an underestimate of rainfall in the east by scattering-based passive microwave retrievals and an overestimate by the ice-sensitive MSU retrieval. While the impact of ice scattering on the MSU algorithm is reduced due to the large satellite footprint, it is apparent that the MSU retrieval produces more rain in the central-east Pacific ITCZ than emission-based retrievals from SSM/I. Differences in cloud water content between the east and west Pacific may also have a significant impact on the MSU retrievals; however, due to insufficient observations of cloud water between the two regions it is impossible to quantify what effect, if any, this may have. Finally, the impact of an east/west Pacific climate regime bias in the rain layer thickness will result in an underestimate in surface rain rate over the east Pacific. Subsequently, regional biases such as those identified over the east Pacific ITCZ can have a substantial impact on interannual variability in tropical-mean rainfall. In particular, the response of tropical-mean rainfall to El Niño varies in both amplitude and sign between current satellite retrievals.

It is apparent, therefore, that rainfall climate regimes are not only regionally dependent, but exhibit seasonal and interannual variability as well. Preliminary comparisons over other regions indicate that variability associated with monsoon cycles as well as intraseasonal variability is also likely to have an impact on the structure of storm systems. Unfortunately, correcting for regional biases in the satellite retrieval algorithms is not possible without ancillary information on the storm system structure from a source such as the TRMM PR due to time-dependent variations like those produced by the 1997/98 El Niño. Because of the effect of changes in east versus west Pacific precipitation on the resulting interannual variability of tropical-mean rainfall, however, it is very likely that the MSU observations used in Soden's (2000) comparison are, as he suggests, inadequate for monitoring climate variability. Clearly, however, all of the rainfall products investigated in this study are susceptible to biases resulting from spatial and temporal changes in rainfall structure.

Acknowledgments

This work was supported by NASA TRMM Grant NAG5-10482. The TRMM data products were provided through the TRMM Science Data and Information System. ECMWF analysis data was provided from the European Centre for Medium-Range Weather Forecasts via the National Center for Atmospheric Research, and NCEP reanalysis was obtained from NOAA's Climate Diagnostics Center. The monthly GPI and SSM/I rainfall estimates were obtained through the Global Precipitation Climatology Project courtesy of George Huffman, and SST and water vapor products from TMI were obtained from Remote Sensing Systems.

REFERENCES

  • Adler, R. F., , C. Kidd, , G. Petty, , M. Morrissey, , and H. M. Goodman, 2001: Intercomparison of global precipitation products: The third Precipitation Intercomparison Project (PIP-3). Bull. Amer. Meteor. Soc., 82 , 13771396.

    • Search Google Scholar
    • Export Citation
  • Arkin, P. A., , and B. N. Meisner, 1987: The relationship between large-scale convective rainfall and cold cloud over the Western Hemisphere during 1982–84. Mon. Wea. Rev., 115 , 297316.

    • Search Google Scholar
    • Export Citation
  • Arkin, P. A., , R. Joyce, , and J. E. Janowiak, 1994: The estimation of global monthly mean rainfall using infrared satellite data: The GOES precipitation index (GPI). Remote Sens. Rev., 11 , 107124.

    • Search Google Scholar
    • Export Citation
  • Berg, W., 2000: Differences in satellite estimates of precipitation over the east Pacific. Preprints, 10th Conf. on Satellite Meteorology and Oceanography, Long Beach, CA, Amer. Meteor. Soc., 375–377.

    • Search Google Scholar
    • Export Citation
  • Cess, R. D., , M. Zhang, , P. H. Wang, , and B. A. Wielicki, 2001: Cloud structure anomalies over the tropical Pacific during the 1997/98 El Niño. Geophys. Res. Lett., 28 , 45474550.

    • Search Google Scholar
    • Export Citation
  • Ferraro, R. R., , and G. F. Marks, 1995: The development of SSM/I rain-rate retrieval algorithms using ground-based radar measurements. J. Atmos. Oceanic Technol., 12 , 755770.

    • Search Google Scholar
    • Export Citation
  • Gruber, A., , X. Su, , M. Kanamitsu, , and J. Schemm, 2000: The comparison of two merged rain gauge–satellite precipitation datasets. Bull. Amer. Meteor. Soc., 81 , 26312644.

    • Search Google Scholar
    • Export Citation
  • Houghton, J. T., , L. G. Meira Filho, , B. A. Callender, , N. Harris, , A. Kattenberg, , and K. Maskell, Eds.,. . 1996: Climate Changes 1995: The Science of Climate Change. Cambridge University Press, 572 pp.

    • Search Google Scholar
    • Export Citation
  • Huffman, G. J., and Coauthors. 1997: The Global Precipitation Climatology Project (GPCP) combined precipitation dataset. Bull. Amer. Meteor. Soc., 78 , 520.

    • 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
  • Jaeger, L., 1983: Monthly and areal patterns of mean global precipitation. Variations in the Global Water Budget, A. Street-Perrott, M. Beran, and R. Ratcliffe, Eds., D. Reidel, 129–140.

    • Search Google Scholar
    • Export Citation
  • Janowiak, J. E., , P. A. Arkin, , P. Xie, , M. L. Morrissey, , and D. R. Legates, 1995: An examination of the east Pacific rainfall distribution. J. Climate, 8 , 28102823.

    • Search Google Scholar
    • Export Citation
  • Kummerow, C., 1993: On the accuracy of the Eddington approximation for radiative transfer in the microwave frequencies. J. Geophys. Res., 98 , 27572765.

    • Search Google Scholar
    • Export Citation
  • Kummerow, C., 1998: Beamfilling errors in passive microwave rainfall retrievals. J. Appl. Meteor., 37 , 356370.

  • Kummerow, C., , W. Barnes, , J. Shiue, , and J. Simpson, 1998: The Tropical Rainfall Measuring Mission (TRMM) sensor package. J. Atmos. Oceanic Technol., 15 , 809817.

    • Search Google Scholar
    • Export Citation
  • Kummerow, C., and Coauthors. 2001: The evolution of the Goddard profiling algorithm (GPROF) for rainfall estimation from passive microwave sensors. J. Appl. Meteor., 40 , 18011820.

    • Search Google Scholar
    • Export Citation
  • Legates, D. R., , and C. J. Wilmott, 1990: Mean seasonal and spatial variability in gauge-corrected, global precipitation. Int. J. Climatol., 10 , 111127.

    • Search Google Scholar
    • Export Citation
  • Petty, G. W., 1994a: Physical retrievals of over-ocean rain rate from multichannel microwave imagery. Part I: Theoretical characteristics of normalized polarization and scattering indices. Meteor. Atmos. Phys., 54 , 7999.

    • Search Google Scholar
    • Export Citation
  • Petty, G. W., 1994b: Physical retrievals of over-ocean rain rate from multichannel microwave imagery. Part II: Algorithm implementation. Meteor. Atmos. Phys., 54 , 101121.

    • Search Google Scholar
    • Export Citation
  • Petty, G. W., 1995: Frequencies and characteristics of global oceanic precipitation from shipboard present-weather reports. Bull. Amer. Meteor. Soc., 76 , 15931616.

    • Search Google Scholar
    • Export Citation
  • Petty, G. W., 1999: Prevalence of precipitation from warm-topped clouds over eastern Asia and the western Pacific. J. Climate, 12 , 220229.

    • 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.

  • Spencer, R. W., 1993: Global oceanic precipitation from the MSU during 1979–91 and comparisons to other climatologies. J. Climate, 6 , 13011326.

    • 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
  • Wentz, F. J., 1997: A well calibrated ocean algorithm for special sensor microwave/imager. J. Geophys. Res., 102 , 87038718.

  • Wentz, F. J., , and R. W. Spencer, 1998: SSM/I rain retrievals within a unified all-weather ocean algorithm. J. Atmos. Sci., 55 , 16131627.

    • Search Google Scholar
    • Export Citation
  • Wentz, F. J., , C. Gentemann, , D. Smith, , and D. Chelton, 2000: Satellite measurements of sea surface temperatures through clouds. Science, 288 , 847850.

    • Search Google Scholar
    • Export Citation
  • Wilheit, T. T., , A. T. C. Chang, , M. S. V. Rao, , E. B. Rodgers, , and J. S. Theon, 1977: A satellite technique for quantitatively mapping rainfall rates over the oceans. J. Appl. Meteor., 16 , 551560.

    • Search Google Scholar
    • Export Citation
  • Wilheit, T. T., , A. T. C. Chang, , and L. Chiu, 1991: Retrieval of monthly rainfall indices from microwave radiometric measurements using a probability distribution function. J. Atmos. Oceanic Technol., 8 , 118136.

    • Search Google Scholar
    • Export Citation
  • Xie, P. P., , and P. A. Arkin, 1997: Global precipitation: A 17-year monthly analysis based on gauge observations, satellite estimates, and numerical model outputs. Bull. Amer. Meteor. Soc., 78 , 25392558.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

Mean DJF oceanic rainfall for the 10-yr period from 1987 to 1996 from available long-term satellite-based rain estimation techniques. The retrievals shown include (a) the GPI based on the technique developed by Arkin and Meisner (1987) and Arkin et al. (1994), (b) Spencer's (1993) estimates from the MSU, and (c) SSM/I emission-based retrievals from Wilheit et al. (1991). Boxes indicating the locations of the selected west (2°–12°N, 130°–160°E) and east Pacific (5°–15°N, 120°–150°W) regions are also shown

Citation: Journal of Climate 15, 24; 10.1175/1520-0442(2002)015<3659:DBEAWP>2.0.CO;2

Fig. 2.
Fig. 2.

Satellite-derived estimates of (a) SST from TMI (Wentz et al. 2000), (b) integrated water vapor from TMI (Wentz and Spencer 1998), (c) TMI rainfall (Kummerow et al. 2001), and (d) PR rainfall (Iguchi et al. 2000) for the 3-month period from DJF 1999/2000

Citation: Journal of Climate 15, 24; 10.1175/1520-0442(2002)015<3659:DBEAWP>2.0.CO;2

Fig. 3.
Fig. 3.

A comparison of mean profiles of (a) temperature, (b) temperature difference, and (c) relative humidity from ECMWF analysis for the selected east and west Pacific regions during DJF 1999/2000. Specified sea surface temperature values are from TMI (Wentz et al. 2000)

Citation: Journal of Climate 15, 24; 10.1175/1520-0442(2002)015<3659:DBEAWP>2.0.CO;2

Fig. 4.
Fig. 4.

Mean rainfall profiles from the TRMM PR for the selected east and west Pacific regions for DJF 1999/2000. Mean profiles are shown for both (a) stratiform and (b) convective rain systems

Citation: Journal of Climate 15, 24; 10.1175/1520-0442(2002)015<3659:DBEAWP>2.0.CO;2

Fig. 5.
Fig. 5.

Histograms of the relative contributions of convective, stratiform, and total rainfall from the TRMM PR for five different categories of surface rainfall rates during DJF of 1999/2000 over the selected east and west Pacific regions. Three pairs of bars corresponding to convective (C), stratiform (S), and total (T) rainfall are shown for each rain-rate category. For each pair the left-hand side bar (dark shading) corresponds to the west Pacific region while the right-hand side bar (light shading) is for the east Pacific. The far right-hand panel shows the totals for all five rain categories combined. Surface rain-rate values below 0.5 mm h−1 are excluded due to noise in the lightest rainfall cases

Citation: Journal of Climate 15, 24; 10.1175/1520-0442(2002)015<3659:DBEAWP>2.0.CO;2

Fig. 6.
Fig. 6.

Two-dimensional histograms of TMI scattering versus emission indices over the (a) west and (b) east Pacific regions. The histogram values include raining pixels only and they have been normalized by the total number of raining pixels in each region. The solid black line is the mean scattering index as a function of the emission index and the two dotted lines are provided for reference

Citation: Journal of Climate 15, 24; 10.1175/1520-0442(2002)015<3659:DBEAWP>2.0.CO;2

Fig. 7.
Fig. 7.

Simulated brightness temperatures as a function of cloud liquid water and rainfall computed using an Eddington radiative transfer model (Kummerow 1993) and a simple six-layer cloud in the Tropics. Results are shown for (a) the 50.3-GHz MSU channel used for the MSU precipitation retrievals (Spencer 1993) and (b) the 19.35-GHz channel, which is the primary emission channel used for many SSM/I and TMI retrievals. The dashed line shows the scattering effect of adding ice to the cloud above the freezing level

Citation: Journal of Climate 15, 24; 10.1175/1520-0442(2002)015<3659:DBEAWP>2.0.CO;2

Fig. 8.
Fig. 8.

Total rainfall (normalized) as a function of storm height over the selected east and west Pacific regions for DJF 1999/2000. The results are divided into separate categories for (a) small, (b) medium, and (c) large storm systems. Total rainfall within each 0.5-km storm height category has been normalized so that the area under the curve is the same for both regions and the maximum value is one

Citation: Journal of Climate 15, 24; 10.1175/1520-0442(2002)015<3659:DBEAWP>2.0.CO;2

Fig. 9.
Fig. 9.

A comparison of the mean DJF 1999/2000 height of the freezing level, or 0°C isotherm minus the height of the brightband or melting layer height from the TRMM PR. Results are shown for estimates of freezing level height from several sources including (a) NCEP reanalysis, (b) ECMWF analysis, and (c) retrievals from TRMM TMI using the technique by Wilheit et al. (1991)

Citation: Journal of Climate 15, 24; 10.1175/1520-0442(2002)015<3659:DBEAWP>2.0.CO;2

Fig. 10.
Fig. 10.

Mean rainfall profiles from the TRMM PR for the selected east and west Pacific regions for DJF 1997/98 during the intense El Niño event. Mean profiles are shown for both (a) stratiform and (b) convective rain systems

Citation: Journal of Climate 15, 24; 10.1175/1520-0442(2002)015<3659:DBEAWP>2.0.CO;2

Fig. 11.
Fig. 11.

Time series of rainfall anomalies from GPI and SSM/I (Wilheit et al. 1991). The mean rainfall values are averaged over the tropical oceans between 30°N and 30°S for (a) the west Pacific (110°E–160°W), (b) the east Pacific (160°–70°W), and (c) the global Tropics. For purposes of comparing the relative variability, the means have been removed and the values scaled by the standard deviation of the time series. A 13-month running mean filter has also been applied to the data

Citation: Journal of Climate 15, 24; 10.1175/1520-0442(2002)015<3659:DBEAWP>2.0.CO;2

Save