1 2 Contribution of extreme convective storms to rainfall in South America 3 4 5 6 7 By K. L. Rasmussen 1, 2 , M. M. Chaplin, M. D. Zuluaga 3 , and R. A. Houze, Jr. 8 9 10 Department of Atmospheric Sciences 11 University of Washington 12 Seattle, WA 13 14 15 Submitted to the Journal of Hydrometeorology 16 April 2015 17 Revised August 2015 18 19 20 1 Corresponding author: Kristen Lani Rasmussen, Department of Atmospheric Sciences, University of Washington, Box 351640, Seattle, WA 98195 E-mail address: [email protected]2 Current affiliation: National Center for Atmospheric Research, Boulder, CO 3 Current affiliation: Universidad Nacional de Colombia, Medellín
44
Embed
Contribution of extreme convective storms to rainfall in ... · " 3" 34" 1. Introduction 35" Precipitation from thunderstorms and mesoscale convective systems (MCSs) greatly 36" influence
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
2
Contribution of extreme convective storms to rainfall in South America 3
4
5
6
7
By K. L. Rasmussen1, 2, M. M. Chaplin, M. D. Zuluaga3, and R. A. Houze, Jr. 8
9
10
Department of Atmospheric Sciences 11
University of Washington 12
Seattle, WA 13
14
15
Submitted to the Journal of Hydrometeorology 16
April 2015 17
Revised August 2015 18
19
20
1 Corresponding author: Kristen Lani Rasmussen, Department of Atmospheric Sciences, University of Washington, Box 351640, Seattle, WA 98195 E-mail address: [email protected] 2 Current affiliation: National Center for Atmospheric Research, Boulder, CO 3 Current affiliation: Universidad Nacional de Colombia, Medellín
2
ABSTRACT 21
The contribution of extreme convective storms to rainfall in South America is 22
investigated using 15 years of high-resolution data from the Tropical Rainfall Measuring Mission 23
(TRMM) Precipitation Radar (PR). Precipitation from three specific types of storms with 24
extreme horizontal and vertical dimensions have been calculated and compared to the 25
climatological rain. The tropical and subtropical regions of South America differ markedly in the 26
influence of storms with extreme dimensions. The tropical regions, especially the Amazon Basin, 27
have aspects similar to oceanic convection. Convection in the subtropical regions, centered on 28
the La Plata Basin, exhibits patterns consistent with storm lifecycles initiating in the foothills of 29
the Andes and growing into larger mesoscale convective systems that propagate to the east. In 30
the La Plata Basin, convective storms with a large horizontal dimension contribute ~44% of the 31
rain and the accumulated influence of all three types of storms with extreme characteristics 32
produce ~95% of the total precipitation in the austral summer. 33
3
1. Introduction 34
Precipitation from thunderstorms and mesoscale convective systems (MCSs) greatly 35
influence agricultural and socioeconomic conditions in South America. Yet these storms often 36
occur in regions without routine ground-based meteorological observations. The launch of 37
satellites with spaceborne radars has made it possible to study the physical characteristics of 38
storms in such regions, and the Tropical Rainfall Measuring Mission (TRMM) satellite with its 39
precipitation radar was in orbit long enough to evaluate these characteristics climatologically. 40
The satellite radar's ability to discern storm structures in three dimensions makes it possible to 41
determine the nature of the precipitating systems producing the rainfall. Using these radar data to 42
study the frequency, intensity, and structure of extreme precipitation events in the current climate 43
will lay groundwork for anticipating changes in these patterns of precipitating convection as 44
climate changes occur. 45
It is important to examine the climatology of storm type as well as rainfall. In general, 46
regions experiencing the most precipitation typically do not coincide with regions known to 47
support storms with the most extreme vertical structure (Zipser et al. 2006). Hence, flash 48
flooding and severe local weather may be favored despite low overall rainfall. On the other hand, 49
mesoscale convective systems (MCSs; Houze 2004), i.e. storms with large horizontal 50
dimensions, can contribute large fractions of warm season rainfall due to their breadth, long-51
lived nature, and repeated occurrence in certain regions (Fritsch et al. 1986; Durkee et al. 2009). 52
It therefore seems important to determine the climatology of convective storms with differing 53
horizontal and vertical scales. 54
The goal of this study is to assess the relative contribution of precipitation from storms 55
with the most extreme horizontal and vertical dimensions to the climatological precipitation in 56
4
South America. Some past studies have focused on the contributions to rainfall by systems of 57
different horizontal dimension (e.g., Liu 2011; Romatschke and Houze 2013). Other studies have 58
examined the statistics of vertical echo structure (e.g. Boccippio et al. 2005). In the present 59
study, we use a technique that uses metrics of both vertical and horizontal dimensions to identify 60
deep convective systems of different types, corresponding to different stages of convective 61
system development. Romatschke and Houze (2010), Rasmussen and Houze (2011), and 62
Rasmussen et al. (2014) have used this methodology to examine climatological patterns of 63
convection across South America. These studies showed how the storms vary regionally in 64
character. However, these studies did not determine the contribution from storms of different 65
types and in different stages of development to the rainfall climatology of the continent. That is 66
the purpose of this paper. 67
2. Data and methodology 68
2.1 Regions of study 69
For reference, Figure 1 shows the regions examined in this study as well as the major 70
rivers in South America. Our methodology identifies the contributions by different storm types in 71
the indicated regions. 72
2.2 Identification of extreme echo features and storm types 73
The TRMM PR has fine three-dimensional spatial resolution (4-5 km horizontal, 250 m 74
vertical) with a near-uniform quasi-global coverage that permits comprehensive analysis between 75
36°N and 36°S (Kummerow et al. 1998; 2000). This study uses 15 years of V7 TRMM PR data 76
from January 1998 to December 2012 for the austral spring through fall seasons. We focus on 77
the austral spring (SON) and summer (DJF) to specifically address the impact of storms with 78
5
extreme characteristics on warm season convective precipitation. Additionally, Houze et al. 79
(2015) demonstrate that in regions experiencing frequent midlatitude frontal systems during the 80
winter, stratiform echoes are more likely to be produced by frontogenetical processes rather than 81
having their origin in convective processes. Thus, we limit the focus of this study to the seasons 82
most likely to experience convective storms and MCSs in South America. The following TRMM 83
data products are used: 84
• 2A23—rain characteristics (Awaka et al. 1997); Rain is separated into three categories: 85
convective, stratiform, and other; all references to convective and stratiform precipitation 86
are based on these classifications 87
• 2A25—rainfall rate and profile (Iguchi et al. 2000); provides the attenuation-corrected 88
three-dimensional reflectivity data 89
These data were processed following the methodology of Houze et al. (2007) and Romatschke et 90
al. (2010). All of the fields were mapped onto a 0.05° by 0.05° latitude-longitude Cartesian grid. 91
We first identify all three-dimensional echo objects that have detectable radar reflectivity 92
consisting of two or more contiguous horizontal pixels. Each such object is defined as a 93
distinguishable rain area (DRA; Houze et al. 2015). 94
All pixels located in DRAs are identified, and the precipitation rates in those pixels are 95
calculated. We search each DRA to determine if it contains embedded within it certain extreme 96
characteristics. The embedded features that we identify here have been defined and used in 97
previous studies of continental convection by Houze et al. (2007), Romatschke and Houze 98
(2010), Romatschke et al. (2010), Houze et al. (2011), Rasmussen and Houze (2011), Rasmussen 99
et al. (2013, 2014, 2015), Zuluaga and Houze (2015), Rasmussen and Houze (2015), and Houze 100
et al. (2015). These embedded features are defined as: 1) deep convective core (DCC) which is a 101
6
three-dimensional contiguous 40-dBZ echo ≥ 10 km in maximum height; 2) wide convective 102
core (WCC) which is a contiguous three-dimensional 40-dBZ echo with a maximum horizontal 103
dimension ≥ 1000 km2 and 3) broad stratiform region (BSR) which is a region of contiguous 104
stratiform echo ≥ 50,000 km2 in horizontal dimension. These embedded echo features have a 105
relationship to the stage of development of the storm producing the echoes. DCCs represent 106
especially intense convection and tend to be found in earlier stages of development. WCCs 107
represent very intense convection that has grown upscale to form a mesoscale unit of intense 108
convection. BSRs occur in the mature and later stages of development of MCSs (Houze 2004). 109
Using ground-based radar, Zuluaga and Houze (2013) demonstrated that these types of echo 110
features indeed represent early, middle and later stages of convective system development in a 111
statistical sense. 112
We define “storm type” according to whether a DRA contains one of the categories of 113
embedded extreme echo features (DCC, WCC, or BSR). If so, then the DRA is referred to as a 114
storm containing a DCC, WCC, or BSR. Then we determine whether or not the rain observed by 115
the TRMM PR at a given location and time was falling from a DRA containing one of the above 116
echo types. The main objective of the paper is to compile rainfall statistics on storms containing 117
DCCs, WCCs, and BSRs. We have also defined a separate category of storms containing both 118
DCCs and WCCs. However, for conciseness we have not included this combined category in 119
some of the maps presented below. Suffice it to say that those mapped patterns lie between those 120
of storms containing DCCs and WCCs, and we will present summary statistics that include this 121
combined category. 122
7
2.3 Precipitation estimation 123
Iguchi et al. (2009) suggested that the V7 TRMM PR 2A25 rainfall algorithm tends to 124
underestimate the precipitation from deep convection over land. Rasmussen et al. (2013) 125
investigated the scope of this bias in extreme storms in South America and confirmed that the V7 126
TRMM PR 2A25 algorithm tends to underestimate rain in all three extreme echo types used in 127
the current study due to insufficiencies in the rain algorithm in capturing the full characteristics 128
of deep convective storms over land regions. Rasmussen et al. (2013) showed that lower 129
estimates by the algorithm are most biased for extreme precipitating systems that contain 130
significant mixed phase and/or frozen hydrometeors. Regions of South America that experience 131
the most frequent storms containing deep convective cores are in the subtropics that do not 132
regularly receive large amounts of climatological rainfall, thus an underestimation of the 133
climatological precipitation can influence the perception of the climatology and hydrologic cycle 134
in South America. 135
To mitigate the TRMM PR algorithm bias for the types of overland storms studied here, 136
we adopted the methodology of Rasmussen et al. (2013) who proposed using the Z-R137
relationship 138
Z = aRb (1) 139
140
where Z is the equivalent radar reflectivity factor (mm6 m-3) and R is the rain rate (mm h-1). 141
This relationship is used to estimate surface precipitation from the TRMM PR attenuation-142
corrected reflectivity data. The lowest nonzero value of Z is used at each data pixel between the 143
surface and 2.5 km above ground level for each precipitation echo, which is similar to the 144
procedure used for the TRMM PR algorithm 2A25 (Rasmussen et al. 2013). The parameters a 145
8
and b are constants depending on rain type (convective, stratiform, or other). Rasmussen et al. 146
(2013) examined multiple values for these parameters. Here we use values used previously by 147
Romatschke and Houze (2011), which give a reasonable estimate of precipitation in the tropics 148
and subtropics (convective: a = 100, b = 1.7; stratiform: a = 200, b = 1.49; other: a = 140, b = 149
1.6). 150
2.4 Calculation of storm-type rain contribution 151
One complication in using the TRMM PR data to develop statistics of rainfall by certain 152
storm types is that the sampling by satellite overpasses is intermittent in time and space (Negri et 153
al. 2002). For statistics to be comparable between locations, we must make an accommodation 154
for the irregular sampling. To make the data comparable, we define a parameter (F) to be applied 155
to every horizontal grid element (pixel): 156
F =
NS
NT
⎛⎝⎜
⎞⎠⎟
(2) 157
where NT is the total number of times a given pixel is sampled by a TRMM overpass, and NS is 158
the number of times that pixel is occupied by a certain storm type (defined in Section 2.2) at the 159
time of an overpass. The contribution C of a given storm type to the rainfall at that pixel, is then 160
calculated according to 161
C = F
RS
RT
⎛⎝⎜
⎞⎠⎟
(3) 162
where RT is the average rain rate (mm h–1) seen in a given pixel over the 15 years of TRMM 163
measurements, and RS is the average rain rate within the subset of those grid elements that are 164
occupied by a given storm type. The resulting values represent a satellite overpass-corrected field 165
of the rain contribution from each storm type to the total precipitation in the study region. By 166
9
using this technique, this study will assess how much of the climatological rain is contributed by 167
each extreme storm type and provide insights into the influence of extreme storm-related 168
precipitation on the hydrologic cycle in various regions of South America. 169
3. Background climatology of precipitation and radar echo characteristics 170
3.1 South American hydroclimate 171
The hydroclimate of South America varies strongly from the tropical to subtropical 172
regions. Tropical South America, largely characterized by the Amazon River basin, has a 173
pronounced annual cycle of precipitation that supports the largest rainforest in the world and 174
contributes 20% of the global river freshwater discharge to oceans (Richey et al. 1989). Seasonal 175
variations of the Intertropical Convergence Zone (ITCZ) and South Atlantic Convergence Zone 176
(SACZ) result in wet and dry seasons in tropical South America in the summer and winter, 177
respectively (Kodama 1992; Nogués-Paegle and Mo 1997; Carvalho et al. 2004). In contrast, 178
subtropical South America, largely characterized by the La Plata River basin, receives much less 179
precipitation than the tropical regions in general, but the precipitation variability in the spring 180
and summer is related to the occurrence of both deep convection and mesoscale convective 181
systems (MCSs) in this region (Velasco and Fritsch 1987; Zipser et al. 2006; Salio et al. 2007; 182
Romatschke and Houze 2011; Rasmussen and Houze 2011; Houze et al. 2015). 183
The Andes mountain range affects the hydroclimate of both tropical and subtropical 184
South America. As one of the largest and longest mountain ranges on Earth, the Andes influence 185
the exchange of moisture and heat from the tropical to subtropical regions primarily through the 186
South American Low-Level Jet (SALLJ; Vera et al. 2006) along the eastern foothills of the 187
10
Andes. Deep convection in subtropical South America is related to the presence of the SALLJ 188
(Salio et al. 2007; Rasmussen and Houze 2011; Rasmussen and Houze 2015). In addition, the 189
relationship between steep sloping terrain and heavy rainfall is particularly important for the 190
susceptibility to both flash and slow-rise river flooding and agricultural sustainability in 191
subtropical South America (Latrubesse and Brea 2009). This investigation of the hydroclimate 192
characteristics in tropical and subtropical South America, specifically focused on the most 193
extreme precipitation elements seen by the TRMM satellite, is important for understanding the 194
hydrometeorology in South America with implications for forecasting, agriculture, human 195
consumption, hydropower, streamflow characteristics in tropical vs. subtropical river basins, and 196
future climate projections. 197
3.2 TRMM radar echo characteristics 198
Figure 2 shows the probability of finding a DRA during the austral spring (SON; Fig. 2a) 199
and the climatology of precipitation generated by these events (Fig. 2b). Although the 200
precipitation maps shown in Figure 2b and throughout this study are created with TRMM PR 201
orbital data (i.e., instantaneous measurements with the TRMM PR over an orbital swath), the 202
spatial patterns of rainfall in these 15-year-long samples are in agreement with climatologies of 203
precipitation relying on continuous and merged multi-sensor measurements (e.g., Huffman et al. 204
2001; Rozante et al. 2010; Liu 2015). In the austral spring, more DRAs tend to occur in the 205
tropical Amazon Basin than the subtropics, particularly near the Andes foothills (Fig. 2a). The 206
precipitation climatology for the spring shows an especially robust rain maximum in southern 207
Brazil and northeastern Argentina, likely related to frequent MCSs in this region. Rasmussen et 208
al. (2014) examined the lightning and severe storm characteristics of storms in the austral spring 209
and showed that maxima in lightning flash rates and flood storm reports are collocated with the 210
11
rain maximum seen in Figure 2b, likely related to storms with both deep and wide intense 211
convective characteristics. 212
Figures 3a and 3b show similar maps for the austral summer (DJF). The contrast between 213
the tropics and subtropics regarding the number of DRAs is much greater than for the austral 214
spring (Fig. 3a). The Amazon Basin has a large amount of climatological rain contributed by 215
many raining events in the region. However, the subtropics receive a substantial amount of 216
rainfall despite the low number of events there compared to the tropics. Similar to the austral 217
spring (Fig. 2), the subtropical rainfall is produced by fewer events than tropical rainfall. 218
Additionally, the subtropical region exhibits a distinct longitudinal shift in the probability of 219
DRAs and the climatological precipitation to the west toward the Andean foothills. A similar 220
shift in extreme storm occurrence and lightning production is shown in Rasmussen et al. (2014), 221
consistent with convective initiation and upscale growth into MCSs in subtropical South 222
America (Romatschke and Houze 2010; Rasmussen and Houze 2011; Rasmussen and Houze 223
2015). 224
Figure 4 presents an overview of the probability of finding the three categories of 225
extreme radar echo structures with the TRMM satellite (defined in Section 2.2) for both the 226
spring and summer. DCCs are notably absent in the tropical Amazon (Fig. 4a, d). WCCs and 227
BSRs do occur in the Amazon region, though not as frequently as in the subtropics. Comparing 228
to Figures 2 and 3 and a recent study examining the variable nature of convection globally from 229
TRMM PR data (Houze et al. 2015), we conclude that the convective elements in this region 230
have a less deep maritime-like character, consistent with Mohr et al. (1999), but nevertheless can 231
aggregate into mesoscale units and overall contribute a large amount of precipitation in the 232
Amazon Basin. This characteristic of convection being weaker but nevertheless productive of 233
12
rainfall verifies the frequently made claim that the Amazon region is climatologically similar to a 234
tropical ocean; sometimes this characterization is referred to as the Amazon region being a 235
"green ocean." The Amazon region has large expanses of open water, a moist and shallow 236
boundary layer, and limited surface temperature variability. Recent studies have demonstrated 237
that mesoscale variations in surface heating are an important factor for convective intensity 238
variability over oceans and continents (Robinson et al. 2011). Additionally, Wall et al. (2014) 239
show that the absence of strong low-level wind shear and convergence in tropical South America 240
contributes to the maritime character of the Amazon. Figure 4 is an objective verification of this 241
oceanic characteristic of Amazon convective precipitation. Houze et al. (2015) show that the lack 242
of DCC echoes, especially in summer (Figure 4d) and the frequent occurrence of BSR echoes 243
(Figure 4f) are characteristics similar to those seen over the tropical oceans. 244
The regions with the highest probability of extreme echo occurrence in the subtopics are 245
collocated with the maximum in precipitation in the spring and summer. Previous studies of 246
subtropical South American convective systems in the austral summer have hypothesized a storm 247
lifecycle where convection initiates along the Andean foothills and sometimes grows upscale 248
into MCSs while propagating east or northeast and then decays into broad stratiform regions 249
farther east (Romatschke and Houze 2010; Matsudo and Salio 2011; Rasmussen and Houze 250
2011). The storm type distributions of DCCs, WCCs, and BSRs in spring are heavily 251
concentrated in southern Brazil and northeastern Argentina, unlike the summer patterns. This 252
difference suggests that a somewhat different canonical storm lifecycle occurs during the austral 253
spring and should be investigated in a future study. However, as will be discussed below in 254
Section 5, an analysis of the diurnal cycle of extreme storm-contributed rainfall in both the 255
13
spring and summer show similar eastward propagation and upscale growth from the late 256
afternoon through early morning, but with decreased magnitudes in the spring. 257
Table 1 shows the overall statistics of the total number of DRAs and TRMM-identified 258
extreme echo cores (defined in Section 2.2) for the austral spring through fall seasons, identified 259
in each region defined by the boxes in Figure 1. In general, more storms containing WCCs were 260
identified in most of the regions. However, the ratio of the number of extreme cores to the total 261
DRAs is very low (Table 2). The highest percentages of extreme storms to total events are 262
WCCs in the La Plata North and South regions (1.8 and 1.9%, respectively) that experience 263
frequent MCSs in the spring and summer. As seen in Table 3 showing the average size of the 264
embedded extreme echoes in various regions in South America, WCCs cover a much larger area 265
than DCCs and thus likely contribute more precipitation. Thus, ~2% of the storms in the La Plata 266
Basin likely make up a large fraction of the climatological rain, echoing the discussion of Figures 267
2 and 3 above. This type of assessment of how much rain each storm type contributes provides 268
crucial information for water management and extreme-storm related flood risks in subtropical 269
La PlataLa PlataLa PlataLa PlataLa PlataBasin SouthBasin SouthBasin SouthBasin SouthBasin SouthBasin SouthBasin South
9000
5000
3000
2000
1000500
0
Height(m)
Amazon
NorthFoothills
La PlataBasinNorth
BrazilianHighlandsAlti
Plano
SierrasdeCordoba
La PlataBasin South Atlantic Ocean
0
10°S
20°S
30°S
40°S40°W50°W60°W70°W80°W 30°W
Regions and rivers
Figure 1. Topographical map of South America showing the Andes mountain range, associated terrain features, and major rivers. Selected regions for this study are outlined in maroon and labeled.
a) Probability of distinguishable rain areas (DRAs) b) Precipitation climatology
40°W50°W60°W70°W80°W 40°W50°W60°W70°W80°W
Eq
10°S
20°S
30°S
Probability
0.15
0.13
0.10
0.08
005
003
0.0
Rain rate (mm
hr -1)
0.40
0.33
0.27
0.20
0.13
0.07
0.00
Figure 2. ( a) Geographical distribution of the probability of finding a distinguishable rain area (DRA ≥ 2 pixels) during the austral spring (SON) from 1998-2012. (b) Precipitation climatology for all DRAs identified in (a). The contour inside the continent represents the 500 m terrain elevation.
a) Probability of distinguishable rain areas (DRAs) b) Precipitation climatology
40°W50°W60°W70°W80°W 40°W50°W60°W70°W80°W
Eq
10°S
20°S
30°S
Probability
0.15
0.13
0.10
0.08
005
003
0.0
Rain rate (mm
hr -1)
0.40
0.33
0.27
0.20
0.13
0.07
0.00
Figure 3. Same as in Fig. 2, but for the austral summer (DJF).
0.0
0.5
1.0
1.5
2.0
2.5
3.0
EQ
10°S
20
30
EQ
10°S
20
30
(a)
(b)
(c)
0.0
0.7
1.3
2.0
2.7
3.3
4.0
0.0
5.0
10.0
15.0
20.0
25.0
30.0
4060 507080°W
Spring (SON) Summer (DJF)D
CC
WC
CB
SR
EQ
10°S
20
30
0.0
0.5
1.0
1.5
2.0
2.5
3.0
(d)
(e)
(f)
0.0
0.7
1.3
2.0
2.7
3.3
4.0
0.0
5.0
10.0
15.0
20.0
25.0
30.0
4060 507080°W
Figure 4. Geographical distribution of the probability of finding an event by each extreme type during the austral spring (SON; left column) and summer (DJF; right column) from 1998-2012. The contour inside the continent represents the 500 m terrain elevation.
EQ
10°S
20
30
EQ
10°S
20
30
4060 507080°W
EQ
10°S
20
30
(a) DCC
(b) WCC
(c) BSR
0.00 0.42 0.83 1.25 1.67 2.08 2.50Rain contribution by storm type (x10-2)
Figure 5. Geographical distribution of the rainfall contribution to the total rain by each storm type during the austral spring (SON) from 1998-2012. The contour inside the continent represents the 500 m terrain elevation.
EQ
10°S
20
30
EQ
10°S
20
30
4060 507080°W
EQ
10°S
20
30
(a) DCC
(b) WCC
(c) BSR
0.00 0.42 0.83 1.25 1.67 2.08 2.50Rain contribution by storm type (x10-2)
Figure 6. Same as in Fig. 5, but for the austral summer (DJF).
!
!""!#""$#""%&'
!"#$%!"&'(")$%!"!*'$
*&"+$%(+"&'(!"*$%(!"&'$$$
(",$%!"('("*$%!"!&')",$%(*"#'-"*$%(!".'
#"-$%&",'()".$%+".',*"#$%,!"+'(&"-$%(,"*'$
*"($%("+'$*!"*$%(*"&'$*.")$%,+"#',("-$%*!",'
("-$%*"!'!."#$%,",'$(,"!$%(#",',("!$%(("-'
-"!$%-"+'(#"!$%*("-',("!$%*."#'(&"&$%+"-'
("+$%(".'(."($%(+".'(!"!$%&,"(',#"+$%,("+'
!",$%!"('("*$%("!'!-"-$%*#"*'&*",$%,,"*'
"#
$%&'
(%
)%
*%
+,-./01223!144565217,891.0!4/079.:57./0;!<=>!
*% *%?% @%A%B%&C
Figure 7. Percentage of the accumulated rainfall contribution from each storm type (indicated by the colors in the legend) to the total accumulated precipitation in each region. Values on the left represent the contribution from the austral spring (SON) and the ones in parentheses are from the austral summer (DJF).
40
30
20
10
0MayJan Mar NovSep Feb AprDecOct MayJan Mar NovSep Feb AprDecOct
Month
Rai
n R
atio
(%)
80
60
40
20
0
North Foothills DCCAmazon DCCNorth Foothills WCCAmazon WCCNorth Foothills BSRAmazon BSR
(a) North Foothills & Amazon (b) La Plata South & Sierras de Córdoba
Sierras de Córdoba DCCLa Plata South DCCSierras de Córdoba WCCLa Plata South WCCSierras de Córdoba BSRLa Plata South BSR
Figure 8. Monthly time series of the accumulated rain contribution, expressed as percentages, from the three storm types (DCC, WCC, and BSR) in the regions of the (a) North Foothills and Amazon, and (b) La Plata South and Sierras de Córdoba.
1100
0800
0500
0200
2300
2000
1700
1400
Loca
l tim
e (h
ours
)
Rai
n co
ntrib
utio
n by
sto
rm ty
pe (x
10-2
)
(a) DCC
(b) WCC
(c) BSR
Longitude
1100
0800
0500
0200
2300
2000
1700
1400
1100
0800
0500
0200
2300
2000
1700
1400-80 -70 -60 -50 -40 -30
0.04
0.04
0.10
0.80
0.72
0.64
0.56
0.48
0.40
0.32
0.24
0.16
0.08
0.00
2.50
2.25
2.00
1.75
1.50
1.25
1.00
0.75
0.50
0.25
0.00
2.00
1.80
1.60
1.40
1.20
1.00
0.80
0.60
0.40
0.20
0.00
Figure 9. Time-longitude diagrams representing the diurnal progression of the contribution to the total rain climatology from storms containing (a) DCCs, (b) WCCs, and (c) BSRs for the austral summer (DJF). The black contour in (a) is the contribution from those events that were classified as DCC only (see text). The diagrams are averaged over a meridional band bounded by 36°S-28°S. The gray line in each plot represents the average topographic relief between 36°S-28°S with a maximum height of 3500 m.
Figure 10. Same as in Figure 9, but for the climatological diurnal cycle of precipitation (mm hr-1) from all distinguishable rain areas (DRAs) in subtropical South America during the austral summer (DJF). The diagram is averaged over a meridional band bounded by 36°S-28°S. The gray line in represents the average topographic relief between 36°S-28°S with a maximum height of 3500 m.