This South America has two major precipitating regions that shaped both the Amazon and the La Plata Basins as readily seen in their topography features (Figure 1). Continental and basin scale water balances are economically relevant to human activities, though more importantly, they determine the complexity of eco systems such as those found in the Amazon and Chaco regions. South America has one of the largest drainage systems with about 12% of all fresh water on the planet. A major contributor to the hydrological cycle in the La Plata basin is the frequent occurrence of mesoscale convective systems (MCSs). [1] compared these over North and South America and concluded the latter region produces stronger, larger and longer lasting events. These systems tend to form during evening hours in spring and summer yielding larger amounts of water [2] . The transport of heat and moisture by the South America low level jet—SALLJ [3] assists in the triggering of events close to the Andes Mountain Range. This occurs most frequently near 25S; southward with variable easterly trajectories in fall and spring; and northward in summer [1] . Precipitation climatology of such systems and others that contribute to the water cycle over the La Plata Basin are important and difficult to obtain from point measurements alone to establish major event regimes, diurnal cycle, seasonality, and variability out to inter-annual periods. This paper focuses on precipitation climatology over the La Plata region from a 5-year period of record. Event-based rainfall statistics include spatial and temporal coherence, frequency and intensity of events, preferred locations and diurnal phase of rainfall triggering; continental and coastal ocean patterns of diurnal, seasonal, interannual variability. The underlying dynamical mechanisms and diurnal variability are inferred and discussed based on spatially resolved and hourly cumulative rainfall estimates. This manuscript presents a brief description of the CMORPH method in Section 2 [4] . Interannual, interseasonal and diurnal spatial distribution, frequency and variability of rainfall systems are analyzed for a 5-year period of record in Section 3. Event characteristics are illustrated in Section 4, diurnal patterns and inferred causes thereof are given in Section 5. A brief summary of major results and conclusions is given in Section 6.

CMORPH (“CPC Morphing technique”; [4] ) provides high spatial (~8 km at the equator) and temporal (half-hourly) analyses of precipitation globally from 60˚N to 60˚S. The method blends together precipitation estimates from passive microwave (PMW) instruments (SSMI, AMSR-E, AMSU, TMI) that are generated from published algorithms. However, since these instruments are from low earth orbit platforms, sampling frequency is a major shortcoming. CMORPH alleviates sampling limitations by using geosynchronous infrared (IR) data as a mechanism to interpolate and “morph” the PMW-derived precipitation estimates yielding spatially and temporally complete precipitation fields every 1/2 hour. CMORPH differs from other combined IR/PMW methods (e.g., TMPA; [5] ; PERSIANN; [6] ) in that it does not use rainfall estimates generated from IR data. Rather, it uses IR to determine the motion of raining systems and applies that motion field to raining systems that have been determined from passive microwave (PMW) data. The hypothesis is that the use of IR data to transport and

interpolate precipitation systems (that have been identified and quantified by PMW data) results in more accurate estimates than those from IR data alone. This notion is borne out by several validation studies (e.g., [7] ). The details of the process are provided by [4] .

Inter-annual and intraseasonal monthly precipitation estimation with CMORPH between 2003 and 2004 is shown in Figure 2. The monthly rainfall fields shown include two major moisture convergence zones. One is the Inter Tropical Convergence Zone (ITCZ) that produces the ascending branch of the Hadley circulation over the Amazon Basin. As suggested by the displacement of this major precipitation region, it migrates from North to South over the Amazon Basin from September to March. The second one located between the extra-tropics and mid-latitudes of South America is linked to baroclinic waves that tend to be stationary in early summer and result in the South Atlantic Convergence Zone. Two other precipitation features are of note. One is a vast region of low rainfall accumulation associated to the descending branch of the Hadley Circulation over Northeast Brazil.

This high-pressure system is a permanent feature across the South Atlantic Ocean, extending eastward to South Africa (not shown). This large divergence area transports moisture equator ward and westward by low-level winds. Zonal winds transport most of the moisture in the Amazon Basin. Secondly there is a region of enhanced precipitation caused by seasonal organized convection centered in the Chaco region within the La Plata Basin. The overall climatology of these transient systems is well established in the literature [11] . Figure 2 indicates that this secondary rainfall maximum is apparent in early fall and spring. The dynamical and thermo-dynamical features of this secondary maximum in South America are the main object of this manuscript. Similar rainfall features have been observed and analyzed in North America [10] , Australia [12] and Africa [13] and were found to be associated with elevated terrain, differential heating, while propagating great distances downstream from the areas where they are formed. It will be seen later that the La Plata organized systems tend to resemble others elsewhere in most cases but the longer duration ones are caused by phase locking between the entrance region of upper level jets (ascendance) and the low level jet that injects high moisture content from the Amazon Basin to fuel organized extra tropical convection [3] .

Figure 3 shows the average hourly rainfall accumulation field for South America in 2003. The diurnal cycle over the Amazon and over the La Plata Basins is the dominant rainfall regions that propagate westward and eastward, respectively. The one over the Amazon starts late in afternoon and progress to the next day while the one over the La Plata starts early in the day.

The relative rainfall accumulation in both regions are comparable but the one in the La Plata is much less frequent being more apparent in early fall and spring. In spite of many other diurnal events, these events propagate systematically, spreading rainfall in space and time through Southern South America. Such propagation renders the organized convection clearer. Other important imbedded features are ordinary convective systems over the Chaco region, but these also fail to propagate under the less favorable dynamic conditions associated with reduced vertical wind shear.

The frequency of hourly rainfall accumulation between September and November in 2003 is seen in Figure 4, the most active season of organized convection in the La Plata Basin in a year of more frequent organized convection. It is more evident in the frequency fields that events over the La Plata basin are less frequent than events over the Amazon. However, the average rainfall rate is higher in the La Plata region. Thus, this potentially more baroclinic environment associated with high moist static energy and wind shear seem them both to play an important role while over the Amazon region equatorial waves are less effective in that sense.

Figure 5 presents the frequency distributions of duration (h), phase speed (m·s⁻¹), and span (km) for all episodes of organized convection in the La Plata Basin between December 2002 and June 2008. There is a greater frequency of smaller episodes of organized convection. The average phase speed is 13.0 m·s⁻¹ for eastward moving episodes. Phase speeds of longer spans can be either small or large. Longer (shorter) spans episodes can propagate very fast (very slow) depending on the underlying dynamic and thermodynamic conditions. The spectrum of phase speeds tend to be narrower (broader) for shorter (longer) durations and spans and, there are episodes of higher (lower) phase speed for longer (shorter) spans and durations.

Phase speeds (not shown) have much greater dependence on the starting longitude than on the starting time. The maximum is located in the Andes Cordillera and the minimum over the South Atlantic Ocean. Systems that start between −40˚ to −50˚W longitudes in the afternoon and evening also have faster phase speeds. The high frequency of front genesis [14] and so intense circulation agrees with the present estimate of episodes of organized convection. Episodes have higher frequency between longitudes −70˚ to 60˚W in the afternoon and evening close to the Andes Cordillera. A maximum at −65˚ longitude late evening is remarkable (Figure 6).

Spans are longer when systems start between −70˚ to −60˚W. The spans tend to be shorter in the upstream direction of propagation given contour conditions at East (Atlantic Ocean) and West (Andes Mountains) that

restrains the organized convection by lower friction and higher static stability, respectively. The frequency distribution of episodes at La Plata Basin episodes (Figure 6) stating longitudes are remarkably higher collocated to Andes Cordillera (18%) and between longitudes −50˚ to −30˚W (>10 average) over South Atlantic Ocean.

Figure 7 shows the frequency distribution of starting times of episodes of organized convection in the La Plata Basins. Frequency distribution of starting times in the La Plata Basin is complex, through the three main peaks at 0100 UTC (6.0%), 0900 UTC (6.5%) and 1900 UTC (5.6%) that coincide with the ones in the Amazon Basin (not shown). Two secondary peaks and the minimum are at 1300 UTC (3.6%), 1500 UTC (4.9%) and

0000 UTC. Episodes of organized convection in the La Plata are modulated by the diurnal cycle of convection, as in the Amazon basin, and modified by the other dynamic and thermodynamic features such as transient baroclinic waves (fronts), topographic induced convection (Andes Cordillera) and frontogenesis (South Atlantic Ocean) that tend to interact with the diurnal cycle in any constructive or destructive phase between them. The most intense and lasting ones are associated to topographic induced convection fed by a low level moisture jet coming from the Amazon Basin [15] .

Figure 8 shows the Hovmöller time-longitude diagram for one-hour rainfall accumulation average in a 48-hour period computed with CMOPRH rainfall estimation in the La Plata domain between 2002 and 2008. The main rainfall pattern in the Hovmöller diagram is the diurnal cycle of convection starting near the Andes Cordillera (−67˚W) around 0000 UTC and propagates eastward in about one day to −46˚W. In general, it rains less at the Cordillera, but as the organized convection propagates eastward, it increases towards the Chaco low elevation region when moist static energy is higher in late afternoon. As organized convection propagates away from the Chaco, topography increases eastward; air temperate and moisture content decrease and so rainfall, in spite of induced vertical motion by the topography. A quasi-stationary pattern is observed between seashore and the Atlantic Ocean associated with the inertial circulation along the coast, especially during summer.

Figure 9 is similar to Figure 8 except for the percentage of time with rainfall during winter, spring, summer

and fall (a, b, c, d), respectively. A rainfall frequency pattern is almost stationary in winter centered in the Chaco region. In spring, the pattern propagates from Andes Cordillera eastward while in summer, rainfall frequency increases and stationary patterns are apparent and caused by stronger diabatic forcing. A similar pattern is observed in fall but with lower frequencies. Organized convection occurs preferably in spring and fall triggered at the Andes late in the evening. They tend to intensify as they move towards a richer moist static in phase with the diurnal cycle of diabatic.

A few episodes of organized convection are presented to illustrate common instances of such systems together with their underlying dynamic and thermodynamics. Figure 10 shows a sequence of five episodes of organized convection between 23 and 29 November 2003. The Hovmöller diagram of one-hour rainfall accumulation average is shown in Figure 10(a). They started near the Andes at −67˚W and ended at −45˚W. As they propagated eastward, the rainfall accumulation increased up to the Chaco region and after decreased.

The Hovmöller diagram of 200-hPa zonal wind anomalies in Figure 10(b) shows a positive anomaly evolving in magnitude and extension starting from the Andes and propagating eastward its phase displacement coincides with the phase of the precipitating systems. The average 200-hPa wind field during the episodes (Figure 10(c)) indicated an upper level jet with its equatorial entrance (ascending motion) over the Chaco region. Concomitantly, the average 925-hPa specific humidity field (Figure 10(d)) varied between 14 g·kg⁻¹ to 22 g·kg⁻¹ over

the upper La Plata basin resulting from intense moisture advection by a low level jet (not shown). The spatial distribution of the rainfall accumulation is shown in Figure 10(e).

These key ingredients of the episodes were vertical shear and moisture advection triggered five propagating organized convective systems, and more importantly, the triggering region was over the Andes and late in the evening. This result is similar to the one found by Carbone at al. (2002) in North America downwind from theRocky Mountains. As this long baroclinic wave moved through, it produced cyclogenesis and precipitation over the Atlantic Ocean (Figure 10(c)).

Shorter-lived episodes of organized convection occurred between 02 and 12 Feb 2004 (Figure 11). The Hovmöller diagram of rainfall (Figure 11(a)) indicates three systems between 02 and 05 Feb 2004 starting about 0000 UTC near the Andes. Stationary convection is seen in the afternoon hours of 05 Feb 2004 associated to the diurnal diabatic heating. Organize systems are concentrate towards the east near and over the Atlantic Ocean between 06 and 12 Feb 2004. The corresponding Hovmöller diagram of 200-hPa zonal wind (Figure 11(b)) shows westerlies associated to weak jet (Figure 11(c)). On the other hand, moisture advection (not shown) was stronger with higher specific humidity further south and near the Andes but lower over the Chaco region resulting from pool of colder air behind a surface front (Figure 11(d)). The estimated total rainfall accumulation was less than the previous episodes (Figure 11(e)). In this instance, the dynamics was somehow weaker under drier boundary layer in the La Plata Basin.

Figure 12 shows another two episodes of organized convection one long and another short with higher rainfall accumulation over the La Plata Basin between 11 and 19 Feb 2004. Hovmöller diagrams of rainfall accumulation (Figure 12(a)) and 200-hPa meridional winds (Figure 12(b)) indicate that the longer system was associated with an upper level wave (Figure 12(c)). The 925 hPa specific humidity field (Figure 12(d)) indicates that the longer last system started about 1200 UTC near the Andes associated to moving trough (Figure 12(b)). This system did not yield as much precipitation (Figure 12(d)) as the first episode shown in Figure 10 under similar lower level moist conditions with less favorable position and timing of the associated upper level jet (Figure 12(c)).

An instance of organized rainfall under an intense upper level jet and dry boundary layer conditions between 01 and 12 Aug 2004 is shown in Figure 13. The Hovmöller diagrams of rainfall accumulation (Figure 13(a)) and 200-hPa zonal wind (Figure 13(b)) show convection starting before crossing the Andes at −80˚W and ending near −70˚W associated to the upper level jet (Figure 13(c)) which propagates eastward and triggers convection over the Chaco region on 06 Aug 2004. It is indeed the extension of the one before the Andes but with not favorable moist conditions given the intrusion of colder and dries air from Southern Argentina (Figure 13(d)). The average rainfall accumulation is small and more significant over the Atlantic Ocean as seen in Figure 13(e).

The last example of organized convection starting near the Chaco region associated with a deep upper level trough between 02 and 07 Oct 2004 (Figure 14) resulting three bands of wide spread precipitation as seen in Figure 14(a). The wave pattern is seen in the 200-hPa meridional wind field of Figure 14(b). The mean 200 hPa wind field during the episode (Figure 14(c)) indicates an entrance region of an intense jet over the Atlantic Ocean and another further Southwest crossing the Andes. Even though an intense low-level jet is present at 850-hPa (Figure 14(d)) injecting moisture and triggering convection at 0000 UTC of 03 to 06 Oct 2004 (Figure 14(a)), they do not propagate in the absence of upper level dynamic support. Figure 14(e) shows the average rainfall accumulation during the episode which was associated with the cold front.

The 10 longest and 10 shortest span systems starting at the Andes late in the evening were identified from the Hovmöller statistics to composite mean fields of 850-hPa and 200-hPa winds, 850-hPa temperature and specific humidity. The fields for long and short span composites are shown in Figure 15 and Figure 16, respectively. The most interesting difference in both situations is the position of the upper level jet. The longest systems are in phase with the diurnal heating, low level moisture advection and the equatorial entrance of the upper level jet. The shortest, on the other hand, the upper level jet is a quarter of wave out of phase with the diurnal cycle. Therefore, vertical shear is an important feature to the longest system that also produces most of the observed precipitation.

Most episodes of organized convection with propagation occurred in spring. Composites of vertical cross sec- tions of zonal and meridional winds, specific humidity, 200-hPa and 850-hPa winds and 850-hPa specific

humidity fields from September to November of 2003 to 2005 were obtained and shown in Figure 17. The mean phase speed of organized convection computed from the Hovmöller in Table 2 is 15 m·s⁻¹ during spring. The vertical profile of zonal winds indicate that it corresponds to 400 hPa in Figure 17(a) which is the overall steering level for organized convection.

The mean position of the upper jet (200 hPa) shown in Figure 17(b) is at −30˚S and such that the transverse upward motion at equatorial entrance is between −20˚S and −25˚S and −80˚W and −65˚W. On the other hand, meridional winds (Figure 17(c)) are from the North throughout the troposphere from −70˚ to −50˚W favorable to transport moist from the moist rich Amazon region. The vertical cross section of specific humidity between the Equator to −40˚S and up to 700 hPa is shown in Figure 17(d). A deep layer of moisture more than 10 g kg⁻¹ goes from the ground level at −35˚S to 750 hPa at the Equator. The main jet at 850 hPa (Figure 17(e)) is just to the east of the maximum in specific humidity (Figure 17(f)). These overall dynamic and thermodynamic conditions are consistent with the higher frequency of organized convection in the La Plata in spring. As the satellite rainfall climatology suggests, similar conditions are also found in April.

Table 1 and Table 2 summarize the main statistics of span, duration, phase speed and recurrence of organized convection annually and seasonally obtained with the Hovmöller diagrams between 2003 and 2007 for westward and eastward propagating systems. Westward propagating systems have shorter span, duration and phase speeds than the eastward ones and correspond to 16% of all organized convection in the La Plata region. These systems

are related to direct circulations, sea breezes, in general. The more important eastward propagating system is responsible for higher rainfall accumulations. The longest spans (>2000 km) are of low frequency but are responsible for most of the rainfall accumulation. Figure 18 shows the average rainfall of system with spans 300 - 700 km, 701 - 1700 km and 1701 and 3700 km or a total of 210, 144 and 82 episodes, respectively. Longer system tends to have higher phase speeds and durations.

These “more important” systems hydrologically speaking tend to be more difficult to predict without adequate ground and upper level measurements and so forecast models are unable to reproduce the correct initial and contour conditions where spatial and temporal resolution of surface data, soundings and others are inadequate. But once convection is detected for instance by satellite IR and WV measurements at the Andes late in the evening under a upper level jet with enough vertical shear, it can be predicted its organization and longevity by the 400 hPa zonal winds, the overall steering level of such long lasting precipitating systems.

Unique satellite-derived rainfall estimation with very high spatial and temporal resolution was used to study the main characteristics of organized convection over the La Plata Basins, an important watershed in South America. Episodes of organized convection have been analyzed by means of time-longitude fields of rainfall accumulation and a technique to compute the starting and ending longitudes and times of them, respective phase speeds, spans

and duration. Basic statistics have been obtained for more 3000 episodes in the La Plata Basin. Spring episodes propagate eastward with average span, duration and phase speed of 565 km, 13 hr and 13 m·s⁻¹. Long span eastward episodes in the La Plata Systems start at Andes Cordillera, and have higher phase speeds and higher rainfall accumulation. In the La Plata transient systems such as cold fronts, front genesis and waves trigger by topography alter the diurnal cycle of convection. The CMORPH data set also allowed a quantitative evaluation of the intraseasonal and interannual variation of rainfall accumulation and interannual variability of diurnal cycle.

This research was sponsored by the National Council for Scientific and Technological Development (CNPq), under grant 304978/2011-6. The authors express their gratitude to their respective organizations for providing opportunities and means for this research. The authors are also thankful to CPC/NCEP for providing CMOPRH datasets. Special thanks to Mrs. Silvana H. Y. Silveira for formatting the Figures. The first author is specially grateful to the National Center for Atmospheric Research, Boulder, Colorado, for the logistic support received during the development of this research work.