1000 VOLUME132MONTHLY WEATHER REVIEWq 2004 American Meteorological SocietyA Numerical Study of a Mesoscale Convective System during TOGA COARE.Part II: OrganizationBADRINATHNAGARAJAN*ANDM. K. YAUDepartment of Atmospheric and Oceanic Sciences, McGill University, Montreal, Quebec, CanadaDA-LINZHANGDepartment of Meteorology, University of Maryland, College Park, College Park, Maryland(Manuscript received 17 April 2003, in final form 17 October 2003)ABSTRACTIn Part I, the authors presented a successful numerical simulation of the life cycle of a warm-pool mesoscaleconvective system (MCS) that occurred on 15 December 1992 during the Tropical Ocean Global AtmosphereCoupled Ocean–Atmosphere Response Experiment. In this study, the simulation results of Part I are diagnosedto investigate the organization of the MCS and the convective onsets that occurred during the growing andmature stages of the MCS.During the life cycle of the MCS, four convective onsets occur in the presence of large-scale ascent, convectiveavailable potential energy (CAPE), and surface potential temperature drop-off (SPTD). It is found that the firstconvective onset is caused by the existence of upward motion, CAPE, and SPTD in the model initial conditions.The second convective onset is regulated by the favorable occurrence of SPTD. The third and fourth convectiveonsets arise from the development of upward motion associated with the westward propagation of the quasi-2-day wave. The four mesoscale precipitation features clustered together to form the MCS in response to theevolution of the vertical motion field.The organization of the MCS is characterized by the presence of a midtropospheric mesovortex situated nearthe position of the first convective onset. Analysis of the relative vorticity (RV) budget indicates that themesovortex originates and intensifies largely from vortex stretching induced by deep convective heating. Adecrease in RV above (below) the mesovortex arises because of the combined effects of the tilting and horizontaladvection terms (the tilting, stretching, and solenoidal terms). Our results suggest that the mesovortex playedlittle role in the subsequent onsets (i.e., second, third, and fourth) and that other warm-pool MCSs occurringnear the transequatorial flow are likely to be associated with mesovortices.1. IntroductionWarm-pool convection is organized on many scales.On the largest and smallest length scales are the su-percloud clusters and individual convective cells. Theintermediate scale is represented by mesoscale convec-tive systems (MCSs). In turn, an MCS is composed ofone or more mesoscale precipitation features (MPFs)that define its mesoscale organization (Tollerud and Es-bensen 1985). Convection over the warm pool has beenstudied using satellite and radar data, but numericalstudies are few in number.Several studies of MPFs using infrared imagery high-* Current affiliation: Abdus Salam International Centre for Theo-retical Physics, Trieste, Italy.Corresponding author address: Dr. Badrinath Nagarajan, AbdusSalam International Centre for Theoretical Physics, PWC section,Strada Costiera 11, Trieste I-34014, Italy.E-mail: [email protected] the clustering behavior of warm-pool convection(e.g., Mapes and Houze 1993; Mapes 1993). Mapes(1993) suggested that the clustering of MPFs cannot beattributed solely to the large-scale flow because of thestrong feedback between deep convection and the large-scale circulation. He showed that compensating verticalmotions associated with deep convection renders themesoscale vicinity favorable for the development of newconvection and may be responsible for the clustering ofMPFs. The work of Johnston (1981) was the first ob-servational study of mesoscale convective vortices(MCVs) occurring in association with midlatitude me-soscale convective complexes (MCCs). Recently,Chong and Bousquet (1999) also reported a near-equa-torial MCV associated with the 13 December 1992 Trop-ical Ocean Global Atmosphere Coupled Ocean–Atmo-sphere Response Experiment (TOGA COARE) MCS,indicating that MCVs may play an important role in theorganization and development of convection (Zhang andFritsch 1987; Bartels and Maddox 1991). Thus large-scale circulation, compensating vertical motions asso-APRIL2004 1001NAGARAJAN ET AL.ciated with deep convection, and the MCV-associatedcirculation are possible mechanisms that organizewarm-pool convection.On the other hand, MPFs appear as entities in radarreflectivity, and they are composed of convective andstratiform precipitation regions during the mature anddissipation stages (Leary and Houze 1979). Rickenbachand Rutledge (1998), using radar measurements, studiedthe organization of the MPFs in terms of the spatialdistribution of precipitation. They concluded that overthe 4-month TOGA COARE intensive observing period(IOP), about 50% of deep convection exhibited linearorganization at length scales of about 100 km. LeMoneet al. (1998) investigated the relation between the ori-entation of warm-pool deep convection and the verticalwind shear at the middle and low levels. They foundthat a strong environmental vertical wind shear at low(mid) levels was associated with a shear-perpendicular(shear parallel) convective band. When both the mid-and low-level shears were strong, the structure of theMPF can be complex, depending on the direction of themidlevel shear with respect to the low-level shear. Spe-cifically, when the midlevel shear was directed oppositeto the low-level shear, secondary convective lines par-allel to the midlevel shear extended rearward from theprimary convective lines. However, when the midlevelshear was perpendicular to the low-level shear, onlyconvective lines perpendicular to the low-level shearwere present. Kingsmill and Houze (1999) analyzed 33TOGA COARE precipitation systems over the 4-monthperiod and concluded that the structure and time evo-lution of the COARE MPFs were in qualitative agree-ment with the conceptual model of convection proposedfor the Global Atmospheric Research Program (GARP)Atlantic Tropical Experiment (GATE) by Leary andHouze (1979).In Part I (Nagarajan et al. 2001, hereafter NYZ01),we presented a successful 16-h simulation of the lifecycle of a class 4 (Yuter et al. 1995) MCS, initializedat 0400 UTC 15 December 1992, using an improvedversion of the Mesoscale Compressible Community(MC2) model (Benoit et al. 1997). The model includedthe solar and infrared radiation schemes (Garand