The Cloud and Rain Liquid Water Characteristics of Different Precipitation Regimes in Brazil

Elsevier Editorial System(tm) for Atmospheric Research Manuscript Draft

Manuscript Number:

Title: The Cloud and Rain Liquid Water Characteristics of Different Precipitation Regimes in Brazil

Article Type: SI: Precipitation Sci Part II

Keywords: Cloud Liquid Water, cloud types, Droplet Size distribution, radar, radiometer

Corresponding Author: Mr. Luiz Augusto Toledo Machado, Ph.D.

Corresponding Author's Institution: INPE - Instituto Nacional de Pesquisas Espaciais

First Author: Alan J Calheiros, Ms.

Order of Authors: Alan J Calheiros, Ms.; Luiz Augusto Toledo Machado, Ph.D.

Abstract: Between 2010 and 2012, the CHUVA project collected information regarding cloud and rain trends in different precipitation regimes in Brazil. CHUVA had four field campaigns, located in the North, Northeast and Southeast regions of Brazil, covering the semi-arid, Amazonas, coastal and mountain regions. The purpose of this study is to present statistics related to the integration of cloud and rain liquid water and the profiles for different cloud types and regimes. The synergy of several instruments allows us to describe the cloud process characteristics and to classify rain events. Microwave radiometer, LIDAR, radar, and disdrometer were employed in this study. The rain type classification was made using vertical profiles of reflectivity (VPR) and polarimetric variables from dual-polarization radar (XPOL). The profiles and integrated cloud liquid water (ILWC) was retrieved with a microwave ground-based radiometer using a neural network. For rainy conditions, the profiles from the liquid water content (LWCR) and their integrated (ILWR) properties were estimated by Micro Rain Radar (MRR) and XPOL VPRs. For non-precipitating clouds, the ILWC values were larger for the sites in Northeast Brazil near the coast than for the other regions. For rainy cases, distinct LWCR profiles and ILWR were observed for different rain classifications and regions with a distinctive rainfall regime. The ILWR for the convective systems show the highest values, followed by stratiform and warm systems. The clouds in the Vale do Paraiba and Belem showed the largest reflectivity in the mixed and glaciated layers, respectively. In contrast, the coastal sites show larger values of cloud and rain liquid water content for non-precipitating and warm clouds. The Vale and Belem clouds present the deepest clouds and larger convective cloud liquid water. Several analyses are presented, describing the cloud process and the differences among the regions.

Cover Letter

Dear Silas Michaelides

Please find attached the manuscript “The Cloud and Rain Liquid Water Characteristics of Different Precipitation Regimes in Brazil” from Alan Calheiros and myself. I would like to submit this manuscript to the special issues - “Perspectives of Precipitation Science II”. I would like to apologize for the late submission.

Best Regards

Luiz Machado

Highlights (for review)

Dear Editor and Reviewers,

This study is one of the first results from the CHUVA project. CHUVA that means rain in Portuguese is a Project to study the Cloud processes of tHe main precipitation systems in Brazil: A contribUtion to cloud resolVing modeling and to the GPM (GlobAl Precipitation Measurement). This study discusses the regional characteristics of the cloud liquid water for different cloud types and regions. The first four CHUVA campaigns measured specifics targets, the first one focus on the satellite precipitation estimation of warm clouds as well the second one in Fortaleza. Fortaleza is also associated to organize cloud clusters from the Ocean, both sites are in Northeast of Brazil. The third campaign was held in Belém, in the mouth of Amazonas river´s. Belém typical rainfall regime at this period are the large squall lines that organize precipitation in different space-time scales. The Fourth Campaign was in a region with intense frequency of thunderstorm, the Paraiba Valley in Southeast Brazil. The relationship between cloud liquid water and thickness, the cloud droplets distribution and the cloud and rain liquid water profiles are evaluated. We found interesting features, the experimental measurements, collected over different rain regimes in Brazil, were very useful in classifying and regionally defining the cloud processes of warm, deep and stratiform clouds.

Highlights

 Differences among different measurement techniques of the same physical parameter are discussed.  The comparison between adiabatic liquid water content obtained by radiosonde and by radiometry was very useful for inferring the importance of the entrainment and coalescence processes.  Regional analyses present the typical warm clouds profiles and integrated values for rain and no rain clouds.  Different brightness band (form and height) were found for each region in stratiform clouds.

Best Regards

Luiz Machado *Manuscript Click here to view linked References

The Cloud and Rain Liquid Water Characteristics of Different Precipitation Regimes in Brazil

Alan J. P. Calheiros and Luiz A. T. Machado

Instituto Nacional de Pesquisas Espaciais, Centro de Previsão de Tempo e Estudos Climáticos, Cachoeira Paulista, SP, Brazil

ABSTRACT

Between 2010 and 2012, the CHUVA project collected information regarding cloud and rain trends in different precipitation regimes in Brazil. CHUVA had four field campaigns, located in the North, Northeast and Southeast regions of Brazil, covering the semi-arid, Amazonas, coastal and mountain regions. The purpose of this study is to present statistics related to the integration of cloud and rain liquid water and the profiles for different cloud types and regimes. The synergy of several instruments allows us to describe the cloud process characteristics and to classify rain events. Microwave radiometer, LiDAR, radar, and disdrometer were employed in this study. The rain type classification was made using vertical profiles of reflectivity (VPR) and polarimetric variables from dual-polarization radar (XPOL). The profiles and integrated cloud liquid

water (ILWC) was retrieved with a microwave ground-based radiometer using a neural

network. For rainy conditions, the profiles from the liquid water content (LWCR) and

their integrated (ILWR) properties were estimated by Micro Rain Radar (MRR) and

XPOL VPRs. For non-precipitating clouds, the ILWC values were larger for the sites in

Northeast Brazil near the coast than for the other regions. For rainy cases, distinct LWCR

profiles and ILWR were observed for different rain classifications and regions with a

distinctive rainfall regime. The ILWR for the convective systems show the highest values, followed by stratiform and warm systems. The clouds in the Vale do Paraiba and Belem showed the largest reflectivity in the mixed and glaciated layers, respectively. In contrast, the coastal sites show larger values of cloud and rain liquid water content for non-precipitating and warm clouds. The Vale and Belem clouds present the deepest clouds and larger convective cloud liquid water. Several analyses are presented, describing the cloud process and the differences among the regions. 1. INTRODUCTION

Clouds cover approximately 67.7% of Earth (Rossow and Shiffer 1999), and it is critical to know the physical properties of clouds to diagnose Earth’s energy and water balance. Atmospheric water is found as vapor (gas phase), cloud and rain liquid water (liquid phase) and different types of ice, such as snow and hail (solid phase) (Rogers and Yau, 1989). The significant variability of hydrometeors is due to the complex atmospheric physical processes that directly impact the weather conditions and climate. For example, the quantity of water in the clouds influences the quantity of latent heat and, consequently, the upward and downward motion within the cloud (Zhao and Carr, 1997). The energy balance is also strongly dependent on the amount of water and ice in the clouds (Crewell and Lörnert, 2003; Zhao and Weng, 2002), which directly influences the climate. However, as mentioned by Löhnertet al. (2001), the lack of information concerning these complex processes, especially with respect to cloud microphysics, has limited the available parameterizations in high-resolution numerical models. Unlike other meteorological parameters, the liquid water content of clouds is not measured operationally, and there is little information about the variability of the average properties. The importance of this knowledge goes beyond forecasting and climate modeling: it is also important for the nowcasting of severe events (Greene and Clark, 1972). According to Pruppacher and Klentt (1997), the cloud liquid water content varies considerably among clouds, from approximately 0.2 gm-3 in the initial stage of cumulus cloud development up to 14 gm-3 during severe storms. Cotton et al. (2011) list a series of characteristics associated with different cloud types, showing that the content of liquid water varies significantly. For example, stratus cloud liquid water presents values of approximately 0.05 to 0.25 gm-3, although cases exist in which these values range up to 0.6 gm-3. This is in agreement with Hogan et al. (2005), based on the synergistic use of many active sensors. Although the maximum found for ordinary cumulus clouds was 1 gm-3. However, Lawson and Blyth (1998) found a large variability. Nonetheless, this value is easily exceeded by systems with large vertical development, such as cumulonimbus, which can have values above 1.5 gm-3.

Atlas et al. (1954) and Donaldson Jr. (1955) were among the first to use active remote sensing to study cloud liquid water content and precipitation. According to Hagen and Yuter (2003), the relationships between radar reflectivity and water content are not as frequently considered as ZR relations (radar reflectivity and rain rate, Michaelides et al., 2010); thus, the authors present several relationships that are applicable to radar measurements. However, the attenuation effects must be considered to correct the error sources associated with the content of liquid water, according to Eccles and Mueller (1971). Recently, Zhao et al. (2013) showed different relationships based on polarimetric variables to estimate the rainwater content, as well as the effects of attenuation on the X-band radar retrieval. Meylek et al. (2005) showed several techniques that allow the estimation of liquid water content, in addition to the synergy between the various co-located equipment. According Ebell et al. (2010) the use of a ground based radiometer may assist in the estimation performed by a cloud radar. The use of passive microwave radiometers to estimate the cloud ILW has been widely applied (Peter and Kampfer, 1992; Liljegren et al., 2001; Ware et al., 2003; Westwater et al., 2005; Mätzler and Morland, 2009; Karmakar et al., 2011). The accuracy of these measurements can achieve 16 gm-2, depending on the microwave channel, calibration and weather conditions (Crewell and Lörnert, 2003). The major difficulty is the partitioning of cloud and rain water content within the same cloud. Based on studies of the polarization difference signal in raindrops, performed by Czekala and Simmer (1998) and Czekala et al. (2001), Saavendra et al. (2012) found mean squared errors of 0.144 mm and 0.052 mm, respectively, for cloud and rain liquid water content during precipitation events using active and passive sensors.

The goal of this study is to determine the water content of precipitating and non- precipitating clouds, using passive and active sensors in several field experiments throughout Brazil during the CHUVA project campaigns, further characterizing the quantities both regionally and by cloud type. This work presents the main differences between the profiles of liquid water content (LWC) and the corresponding integrated liquid water (ILW) for the various regimes of precipitation over the continental and coastal regions in tropical or subtropical latitudes and their differences with respect to the cloud types observed. This study discusses the different results obtained with the different sensors and the limitations and errors associated with each type of measurement.

2. DATA AND METHODS

2.1 The CHUVA project The CHUVA project [Cloud process of the main precipitation system in Brazil: A contribution to cloud resolving modeling and the GPM (Global Precipitation Measurement)] is one of the most important experiments focused on understanding the radiative and microphysical processes of continental clouds over Brazil. Field experiments were conducted in different places with different weather patterns, using the same measuring strategy and instruments to study the precipitation regimes throughout the country. During the experiments, polarimetric and vertically pointing radars, LiDAR, microwave radiometers, disdrometers, GPS, radiosondes and various other instruments were used. The purpose of CHUVA is to advance the understanding of cloud processes, mainly the warm cloud, by studying their physical processes and evolution throughout the life cycle of precipitation systems, evaluating and adjusting the precipitation estimation models, and studying the lightning formation processes. One of the main objectives is to minimize the uncertainties in rainfall estimation.

2.2 Measurement Strategy

The data used in this study were obtained from field experiments, conducted between March, 2010, and December, 2011, over various regions of Brazil. Two experiments focused on the characterization of warm clouds, the first of which was performed in Alcantara, MA, from March 3 to April 15, 2010, and the second, held in Fortaleza, CE, from April 4 to May 1, 2011. Both are located on coast of Northeast Brazil. The third field experiment was conducted during the month of June in Belem, PA, in the Northern region of Brazil. This experiment was characterized by the presence of large convective activity, associated with the intrusion of squall lines and the action of the Intertropical Convergence Zone (ITCZ). Between November, 2011, and March, 2012, the CHUVA project was performed in Southeastern Brazil, or more precisely in the Vale do Paraíba, SP, which has a rainfall regime that is associated with storms, local convection and frontal systems. All of the CHUVA project experiments were conducted with similar measurements strategies, as shown in Figure 1, which shows a schematic representation of the site distribution and geographical position of each field experiment. In all of the CHUVA Project experiments, the equipment strategy for each site was distributed as follows.

 The main site: Equipment to characterize the clouds and precipitation was installed here. High-spatial- and temporal-resolution equipment was used to measure the cloud properties and surface rainfall (e.g., disdrometers, rain gauges, weather stations, and radiometers) as well as the vertical distribution (e.g., LiDAR and Micro Rain Radar).  The radar site: Polarimetric radar was installed at this site. Two scan strategies were performed, including a volume scan and Height Indicator Range (RHI). The latter was oriented over the main site.  The radiosonde Sites: At least three sites were used to characterize the thermodynamic operating systems every 6 hours (00, 06, 12, 18 GMT), except for Alcantara where only one site was used.

2.3 Instrument and Limitations

To analyze the liquid water content in the non-precipitating clouds, a ground-based MP3000A radiometer was used (Radiometrics Corp., Ware et al., 2003). This instrument measures the passive radiation at microwave wavelengths in 35 channels, ranging from 22.00 to 30.00 GHz (21 channels), associated with the emission by water vapor, and from 51.00 to 59.00 GHz (14 channels), associated with the emission by oxygen molecules. Details regarding the physical principles can be found in Westwater et al. (1993 and 2005). The MP3000A is a robust instrument designed to handle the most diverse weather conditions. The most sensitive structural part of the instrument is the radome (Rose et al., 2005) through which the radiation passes before reaching the receiver. Problems associated with precipitation and condensation on the radome may produce erroneous estimations of atmospheric parameters. Even when the radome is built with a hydrophobic material and fitted with a Superblower, the presence of water during precipitation events influences the signal. This generates unreliable measurements. Thus, only measurements in moments without rain were considered. Another uncertainty in the measurements performed by radiometers is related to the errors in instrument calibration (Skou and Vine, 2006). To avoid these problems, calibration was performed using liquid nitrogen (Hardy, 1973) before each campaign, and during the measurements, tip calibration was applied (Han and Westwater, 2000; Cimini et al., 2003). Other uncertainties associated with retrieval are described by Hewison (2006). For cases associated with rain clouds, the rain integrated liquid water content (ILWR) and the rain liquid water (LWCR) content profiles were determined using the vertical profiles of reflectivity (VPR) of two radar systems, Micro Rain Radar (MRR, Peters et al., 2005; Leueberger, 2009) and mobile radar (Selex Meteor 50DX for all sites except Alcantara, where a fixed EEC X Band Dual Polarization was employed), including X-band and dual-polarization at 9.365 GHz (XPOL). Both radar systems suffer attenuation effects as a result of rain (see Peters et al., 2010; Doviak and Zrnic, 1993), which can affect and represent the differences in the estimates of LWCR/ILWR between them. The measurement strategy covered the distance from the main site to the radar site, which did not exceed 22 km, to minimize the effects of attenuation. Additionally, attenuation corrections were applied to both radars. Nevertheless, depending on the intensity of precipitation, attenuation can be very strong, particularly for MRR. Moreover, the updrafts and downdrafts can cause significant variations in the droplet distribution estimations, which is directly reflected in the reflectivity determination of MRR (Peters et al. 2005). Hence, the presence of deep convective systems can cause erroneous liquid water content in MRR measurements; therefore, MRR was not used for this purpose.

2.4. The Liquid Water Calculation

During the CHUVA Project experiments, the MP3000A performed continuous thermodynamic soundings in all weather conditions with a temporal resolution less than 2.5 minutes. This measurement provides the temperature (K), relative humidity (%), and liquid water (g/m3) profiles up to 10 km in height, as well as the water vapor (mm) and liquid water (mm) integrations. The retrieval of these parameters was performed using neural networks (Solheim et al. 1998). Historical radiosondes data sets from locations near the sites of each campaign were used for network training. The retrieval for the cloud integrated liquid water (ILWC) content by ground-based radiometer was made only for conditions without rain. According to Won et al. (2009), the formation of rain in clouds can influence the signals received by ground-based radiometers for up to 2 hours before the rain begins, while the greatest difference between microwave channels can be observed within 30 minutes prior to the rainy event. Therefore, for the analysis of the non-precipitating clouds, ILWC values were not estimated in the period of 30 minutes before and after the rain so that the raindrop presence in the clouds would not influence the ILWC estimation. Such rainy events were classified from the measurements made by disdrometers, which were installed at each main site. Further, to reduce the effects on the MP3000A estimations of large amounts of raindrops in the non-precipitating clouds, only events in which the maximum reflectivity of VPR, as measured by XPOL radar, was less than 20 dBZ were considered. According to Marshall and Palmer (1948) and Marshall and Gunn (1952), such reflectivity values are associated with a droplet distribution that can cause a rain rate of approximately 0.25 mm/h. Further, detailed analysis of the data allowed the elimination of peak instances associated with estimation errors of the cloud base by radiometer in addition to the periods in which dew had formed on the radome. For radiosondes, the methodology applied by Ingold et al. (1998) was used to calculate the values of adiabatic LWCadia and the integrations (ILWadia). These values were calculated only for non-precipitating events. The classification used to determine the events without rain was based only on the measurements provided by the disdrometer at each main site, where only the cases without precipitation events for 3 hours before and after the sounding launch were used.

The characterization of the total liquid water content in a given cloud at any given can only be performed by using several aircraft, which must simultaneously profile the cloud. It is not possible to specify the mixed layer thickness of ice and water or the densities through indirect measurement. Thus, we chose to perform only two distinct analyses, the non-precipitating event mentioned above and another for clouds with rain. In the latter, only the amount of liquid water in the warm part of the cloud was considered, i.e., the cloud base, measured based on the lifting condensation level (LCL) to a value below the melting level. This approach was selected due to the bright band (BB) effect that occurs in stratiform clouds, which is characterized by a peak of reflectivity that can provide an erroneous LWCR estimation. The sudden change of reflectivity in this layer is associated with changes in the refractive index with respect to the thickness of the water film around the melting hydrometeor (Battan, 1973; Houze,

1996). Thus, the ILWR is calculated by integrating the LWCR of the LCL up to a height of 1 km below the 0°C mean level of each experiment. This layer is considered to be the layer that can potentially transform into precipitation. To determine the average freezing level and LCLs, we used the nearest, both spatially and temporally, radiosonde with respect to each VPR.

The estimated LWCR by XPOL is based on VPR, which in turn is determined by the mean reflectivity of 500 m around the main site with a vertical resolution of 200 m.

Thus, the XPOL LWCR was estimated using the methodology of Greene and Clark (1972). In addition to removing the BB effect, other corrections must be performed, such as those associated with strong signal attenuation by rain. In the MRR correction, a path-integrated attenuation was used (PIA), according to Peters et al. (2010). The XPOL correction methodology is described in Schneebeli et al. (2012). For Alcantara, as different radar systems were employed, a specific bias adjustment was applied using a co-located MRR radar. Further, radar systems can present surface noise effects, such as radio frequency. Therefore, to avoid the noise associated with ILWR values, for both radar types, the threshold must exceed MRRNOISE=0.004 mm and XPOLNOISE=0.025 mm. These thresholds are based on the observed distributions in which the LiDAR identified the absence of clouds. The rainfall events for each site were identified as the events whose precipitation, as measured by disdrometer, was greater than 0.1 mm/h.

4. RESULTS

4.1 Cloud Liquid Water

Figure 2 shows the relative frequency histograms of ILWC (a) and ILWadia (b) for all CHUVA project experiments used in this work, the statistics of which can be viewed in Table 1. As shown in Figure 2, the majority of the distributions are relatively similar, except for Alcantara, which is characterized by proportionally more cloud liquid water content, as measured by the radiometer. It is our belief that the adiabatic method is quite realistic. In principle, the entrainment effect reduces the liquid water content more than that estimated by adiabatic processes. However, the microphysical (in particular, coalescence) and non-adiabatic effects tend to generate higher contents of liquid water. For Alcantara and Fortaleza, which are the sites located closest to the coast, it was noted that the last two effects dominate, due to the higher measures of liquid water content in the clouds, which is particular true for Alcantara.

Table 1 shows the mean values, standard deviations and estimation methodology for each region. At Alcantara, the mean value for ILWC is 0.36 mm, which is the largest value obtained from the sites. According to Lohnert and Crewell (2003), an ILWC above 0.4 mm may be associated with the existence of raindrops that strongly affect the microwave brightness temperatures. During the experiment, the presence of virga was observed, despite the fact that the site’s thermodynamic properties did not promote precipitation. This shows that the presence of warm clouds over this site led to rapid droplet formation processes, generating clouds with high values of ILWC. It is evident in Figure 2a that, above 0.2 mm, large populations of Alcantara clouds exist. However, in the adiabatic distributions (Figure 2b), not co-located with the radiometer, the frequency of values above 0.6 mm is significantly lower in comparison to the radiometer estimates. Thus, the microphysics and non-adiabatic processes are emphasized as the predominant processes in Alcantara in particular, although they also occur in the other sites. Vale do Paraiba has the lowest liquid water content, while Belem and Fortaleza have similar values. As previously mentioned, the non-precipitating clouds in Alcantara, on average, have values that are approximately 2.5 times larger than that of the Vale. Although Belem and Fortaleza presented similar mean values (approximately 0.23 mm), Belem had a number of events without rain, which was significantly fewer than that of Fortaleza. This finding suggests that, when clouds are present in Belem, they quickly develop into deep convection clouds, which are associated with precipitation. Generally, Table 1 shows that the average behaviors of the adiabatic estimates are higher than that estimated by the radiometer for the more continental sites. This finding stresses the importance of the entrainment process of the non-saturated air within clouds causing the evaporation of droplets and the consequent reduction of the LWC values in these regions.

The Fortaleza experiment LiDAR and ground-based radiometer systems were co- located to analyze the relationship between the cloud thickness and the non- precipitating liquid water content. Figure 3 shows boxplots with the values of liquid water for each cloud layer, in ranges of 100 meters. After analyzing this relation, we note that the median ILWC increases with cloud thickness. Further, it is evident from the figure that there is greater variability in the values obtained for thicker clouds. It should be noted that only layers below 300 m were observed. Korolev et al. (2007), working with ice stratus clouds, found a linear relationship between ILW and cloud thickness. In the present study, the clouds are non-rainy and are significantly thinner than those studied by Korolev et al.

4.2 Rain Liquid Water

4.2.1. Integrated Liquid Water

This section analyzes the distribution of rain liquid water content, integrated up to 1 km below the freezing level. Figure 4 shows the distribution of the ILWR measurements, estimated by MRR (a) and XPOL (b), for the CHUVA sites. However, at Alcantara, MRR was not used in the vertical orientation; instead, it was employed at a slant, pointed toward the XPOL radar. The statistics for the different combinations discussed here also can be observed in Table 1. In this case, to generate the values of Table 1, we applied a filter, limiting the estimate of LWCR for the LWC maximum observed on the disdrometers during each campaign to avoid unrealistic values. ILWR values above 1.5 mm were observed and were included in the statistical computations, although they are not shown in Figure 4 due to their low frequency in comparison to the lower values. Upon the inspection of both figures, the rainwater content in the analyzed layer is higher than estimated by XPOL when compared to the MRR data for most of the sites. This result is expected, given the larger attenuation at 24 GHz MRR than 9.36 GHz in XPOL. Further, the XPOL measurement strategy favors the upper layers, in comparison to MRR.

Table 1 shows that the Fortaleza site yielded the most significant difference between the

MRR and XPOL data for the ILWR estimation. This may be due to the presence of larger drops in the analyzed layer, which also has the highest attenuation signal. However, it is important to recall that such measurements were not obtained for Alcantara; therefore, this result cannot be evaluated in this region. Nonetheless, the smallest difference was observed in the Vale, which possibly could be associated with greater numbers of smaller raindrops. Further details on the drop size distribution are given in later sections. Generally, Figure 4a shows that the ILWR distributions are similar among regions based on MRR, which may be the result of attenuation, as mentioned previously. However, further details can be observed by XPOL. Figure 4b shows that the highest values were associated with Fortaleza, which had an average of 0.58 mm, followed by Alcantara at 0.47 mm. These high values may be associated with the conditions of environments that are rich in hygroscopic aerosols, which could contribute to the efficient development of convective rain formation in these regions.

The ILWR values in Belem and Vale are lower than the values of the other sites. This could be linked to the process of cold cloud formation because both regions are characterized by strong convective activity, as evidenced by higher rain rates (above 60 mm/h) (see Table II), which are related to deep convective clouds.

In addition to the ILWR values in the liquid cloud layer, the VIL for the mixture (0 to - 20 °C) and glaciated (-20 to -40 °C) layers were also calculated. The events observed in the Vale do Paraiba and Belem regions (Table 1) yielded the highest values. Because the convection over this region is intense, where the presence of squall lines and other convective systems are observed, updrafts can cause the rise of super-cooled water in higher layers, which may increase the VIL values significantly. The Vale do Paraiba is a typical example, where the mixed layer is superior to all other regions, and Belem has the highest values in the glaciated cloud layer. From these results, we can conclude that, in Fortaleza and Alcantara, the layers below zero degrees in temperature have higher liquid water contents than the more continental sites, most likely because larger raindrops are present. However, the clouds in Belem and Vale have more developed ice stages and higher rain rates. In comparison, Vale has a more developed mixed layer, while Belem has a more developed glaciated layer.

4.2.2 The Integrated Liquid Water contents of the different cloud types.

This section analyzes the average behaviors and distributions that are associated with the different types of precipitation. The classifications of deep convective, stratiform with bright band and warm clouds are based on information from the XPOL RHI profiles and the local radiosondes. The first classification level is associated with the identification of warm rain events for which the whole profile of reflectivity must be below 0°C. The next step is to identify the bright band in stratiform events, as described above, based on the work of Fabry and Zawadzky (1995). The classification is applied as a function of the vertical variation of radar reflectivity (dZ/dh, dBZ/km) and the polarimetric ρHV variable (the co-polar correlation coefficient), described by Zrnic et al. (1994). If the BB is not found and the reflectivity is greater than the 39-dBZ threshold, then the system is classified as convective (Awaka et al., 2007). Clouds not filling these conditions were not classified.

Table 1 shows that, while warm and stratiform clouds had similar liquid water contents, the convective events were significantly higher. The ILWR distributions for the three classes and for each site can be observed in Figure 5. Because the histogram associated with the warm clouds shows larger values than that of the stratiform clouds for a few cases, the charts were limited to 2.5 mm. It can be observed in Figure 5a,b that the values above 0.6 mm occur more frequently in warm than stratiform clouds. However, the smallest values were also observed for the warm clouds. This explains the similarities observed between the mean values of the general classification, presented in Table 1 (the final lines). Table 2 shows the frequency of events classified by each VPR, related to rain rates greater than or equal to 0.1 mm/h and associated with different types of clouds (convective, stratiform with bright band and warm); thus, the mean rain rate is also considered. Note that Table 2 refers only to the specific cloud types described before, clouds such as multilayers, high clouds, stratiform clouds without bright band, and weak convective clouds (Z<39 dBZ) were not considered. It can be observed that Vale and Belem have the largest rainfall intensity for convective clouds and that Alcantara has the largest for the warm and stratiform clouds. The reason for this behavior is discussed further. Note that, in Figure 5, the observed events varied significantly by region, where the mean values of the ILWR can be observed in Table 1. For the sites located near the coast, the liquid water content that was associated with warm clouds (Figure 5a) was higher than the contents observed in Belem and Vale do

Paraiba, where the frequency of events with ILWR below 0.5 mm was greater. However, with respect to stratiform clouds, as observed in Figure 5b, the values were similar to the Vale do Paraiba, Belem and Fortaleza regions. Even so, the largest estimated values were observed in Alcantara, where the stratiform clouds had a more active liquid layer, as presented subsequently. The high liquid water content in the liquid layer classified as stratiform could be associated with the active warm clouds. Active warm clouds can grow to a few meters above the freezing level, thereby creating a thin layer of ice. However, the observed rainfall is associated with rain processes in warm clouds. The convective events presented in Figure 5c show the highest liquid water contents, particularly in Fortaleza, where several severe events with ILWR greater than 8 mm were observed. Belem and Vale had similar distributions because the processes of cold clouds were prevalent in these regions. The convective clouds over the Vale were observed to have high precipitation efficiency; that is, even with the lowest ILWR content, these clouds generated the highest rain rates (see Table 2). Alcantara, in turn, showed the lowest liquid water content, despite also exhibiting the largest liquid water content of non-rainy clouds. This is most likely due to the inhibition of deep convective clouds, which was observed a few times. Moreover, the convective clouds have specific maximums, which must be associated with the occurrence of different droplet size distributions in each region, as discussed in the next section. Generally, only in

Fortaleza did the ILWR measurements for the warm clouds have larger values than that of the stratiform clouds. In Alcantara, the values were slightly smaller, and in Belem and Vale, the ILWR was significantly lower for the warm system than the stratiform system, which was particularly true in Vale. For the latter, this result could be related to the formation of young cells, which are associated with storm propagation over the site.

4.2.3. Vertical LWCR Profiles for the Different Cloud Types

In the previous section, we determined that an intrinsic relationship exists between the distributions of rain liquid water and the different types of precipitating events observed in each region. Our analysis of the vertical profiles of these distributions showed which layer is more important within the analyzed systems, as well as how that layer varies. This is an important parameter for modeling radiative processes and is useful in satellite -3 meteorology. Figure 6 shows the mean profiles of LWCR (gm ) for each site, as estimated by XPOL for warm, stratiform and convective clouds. The MRR profiles were not displayed due to the effect of rainfall attenuation, which was associated with convective events; in such cases, unrealistic results are obtained for the upper levels. Further, the figure shows that the warm events that have the lowest liquid water content, followed by stratiform events, which occur in the layer between the LCL and 0°C-1 km in height. The effect of the bright band, which is disregarded in this study, is evident in the figure in addition to the large contents of liquid water observed in the convective clouds, as expected.

Figure 6a shows a significant difference between the coastal warm cloud profiles, in which significantly more liquid water was observed, and the continental warm cloud profiles. Further, the mean profiles of the stratiform clouds (Figure 6b) show that the bright band height observed near the coast is higher than that of the continental clouds. The mean freezing level is 4.9 km at Fortaleza and Alcantara, 4.5 km at Belem and 4.4 km at the Vale do Paraiba. The warm processes in the coastal clouds would benefit from the presence of a larger layer at the coast than in the continental systems, which would consequently provide a comparatively longer time for droplet development, yielding larger droplets. Additionally, it is important to consider that both of the coastal sites are in regions predominated by subsidence. The subsidence reduces the deep convective process, increasing the lifetimes of the non-raining clouds and the warm cloud processes. As already mentioned, a significant amount of liquid water was observed in the liquid layers in the stratiform clouds at Alcantara. It is possible that some of these clouds were not stratiform but, instead, were warm clouds exhibiting the strong development of a liquid phase and the formation of a small ice cap on top, causing them to be classified as stratiform. There are also similarities between the events observed at the Vale do Paraiba and Belem, as well as in the clouds found in the Alcantara and Fortaleza regions. However, if we look at the rain rates of the stratiform clouds over the sites, as detailed in Table 2, we note that higher values can be observed for Vale than Belem. The same is true for Alcantara in comparison to Fortaleza. This could be the reason that the biggest raindrops were observed in Vale and Alcantara, as noted in the next section. With respect to convective clouds (Figure 6c), Belem showed events with higher LWCR in the glaciated layer, while Vale presented higher values in the mixed layer. The highest values at Fortaleza occurred in the liquid layer, especially at -3 elevations below 3 km, where the average LWCR reached 1.8 gm . The brightness bands were higher at Alcantara and Fortaleza, which can be associated with the aloft ice structure. Because these sites also presented clouds with less deep development, the higher brightness band can be related to the rapid melting of ice into low density ice clouds above the melting layer.

4.2.4. Surface Raindrop Size Distribution

Understanding the raindrop size distribution (DSD) for the different systems over the sites is essential for rain characterization as well as for our understanding of the processes and differences between all of the previously analyzed distributions (Hu and Srivastava, 1995; Martin et al., 2010). DSD varies with precipitation intensity, cloud type, location and measurement instrument (Tokay and Short, 1996; Tokay et al., 2001; Caracciolo et al., 2006; Islam et al., 2012). Figures 7a,b,c show the average concentrations (mm-6m-3) of raindrops with respect to raindrop diameter (mm), as observed by PARSIVEL disdrometer for all sites and rain classifications. These distributions confirm the previous results. With respect to warm clouds (Figure 7a), Alcantara exhibited the largest droplets, while Vale had the highest concentration of smaller raindrops, with these two locations demonstrating the highest and lowest rain rates, respectively. The warm clouds in Belem show drops up to 4.5 mm in diameter, while Fortaleza yielded a maximum of approximately 4 mm, and at Alcantara, the maximum was nearly 8.5 mm. Moreover, in Vale, the events with frequencies similar to that of Fortaleza (Table 2) also exhibited significantly lower rain rates. This finding indicates that the warm cloud layer near the coast has a higher rain liquid water content and is more efficient at raindrop formation. The distributions for the stratiform clouds (Figure 7b) are usually similar. However, peculiarities can be observed in the Alcantara measurements. Where the systems showed the highest contents of liquid water, higher concentrations of drops between 1 and 2.5 mm in size were observed in comparison to that of other sites; additionally, diameters up to 4.5 mm were noted. This feature could be related to the warm clouds, which develop their dynamics in lower layers but retain ice covers above the freezing level, as already discussed. The biggest raindrop in this classification was observed at Vale (5.5 mm). With respect to the convective DSDs (Figure 7c), the largest droplets and concentrations formed by the intense systems of the Vale and Belem are evident and are less significant in Alcantara. However, Fortaleza showed a higher concentration of drops larger than 6 mm, which explains the high values of liquid water observed in the previous analyses. The distributions in Belem and Vale presented similarities up to 5.5 mm. The Vale clouds were followed by Fortaleza clouds (in order of decreasing size); at the first site, the concentration of larger drops was slightly higher, reflecting the organization of the convective clouds generated over Southeast Brazil.

The contribution of each class of raindrop to the total liquid water content was computed using the values of LWCDi (the LWC only for a particular drop, Di) and the total LWC observed for each DSD. Figures 7c,d,e show the diameter contribution (%) for the LWC, based on each classification. Figure 7c shows that the warm events in Vale do Paraiba are more influenced by raindrops smaller than 1 mm. In contrast, the events in Fortaleza are more highly associated with drops between 1 and 2 mm in diameter. Belem and Alcantara showed similar behaviors up to diameters of 3 mm, where the maximum was associated with drops measuring approximately 1.5 mm, which is the same size as that for Fortaleza. However, it was observed that larger raindrops have higher contributions to the total liquid water content in Belem and Vale than in Fortaleza. Among the sites studied, Alcantara presented the most significant contribution of larger drops in the total liquid water content of warm clouds, which confirms the previous analyses. Stratiform clouds (Figure 7d) showed a similar distribution. However, there was a greater contribution of drops below 1.5 mm in size at Vale and Fortaleza, while above this diameter, the greatest contributions were obtained at Belem and Alcantara. Unlike what was observed for the integrated liquid water content, the results from Belem and Alcantara exhibit similar characteristics, with Vale and Fortaleza being more similar to each other. The convective clouds (Figure 7e) observed in Alcantara received greater contributions from raindrops measuring 2 and 3 mm, while at the other sites, larger raindrops were the most important. This result was particularly true at Vale, where the highest rain rates were observed.

4.2.5. The precipitation system and mean ILWR

Understanding the average behavior of the rain liquid water content according to its intensity is essential to determine relationships that can be applied for improvements in satellite precipitation estimations and numerical weather forecasting models. Figure 8 shows the XPOL estimated ILWR values for the events, classified according to their precipitation intensity, i.e., light (1 to 2.5 mm/h), moderate (2.5 to 10 mm/h), intense (10 to 50 mm/h) and severe rain (more than 50 mm/h). The values tend to increase gradually with the precipitation rate, and in some cases, this increase is more pronounced, such as in Fortaleza. Our analyses of the rain intensity show as the rain rate increases, the liquid water content also increases, specifically within the rain layer up to a 1 km below the 0°C elevation. It is also worth noting that, in the light to moderate event classifications, a growth of the mean ILWR values can be observed, even if small. This slope is steeper from the moderate to intense rainfalls, and increasing with rainfall next class, which further increases the regional variability. As shown by our results, both the regional and cloud type variability are significant, depending on several features, such as drop size and concentration, cloud layer and the ice processes. Table 2 and Figure 7 show that the raindrop formation processes result in different rain rates, which are related to the variations in DSDs, as shown in the previous section. Therefore, it is not possible to define a universal relationship between ILW and RR. Each region and cloud type has a specific behavior. Based on our results, for the warm processes, the thermodynamic properties (deep convection inhibition) and aerosols (rich coastal aerosol) should be taken into account. For deep convective clouds, this relationship should consider the processes that occur in the mixed and glaciated layers. Meanwhile, for stratiform clouds, this relationship appears to be significantly more universal, except for the fact that some clouds appear to be classified as stratiform but exhibit distinct behaviors (i.e., warm clouds with ice caps).

5. CONCLUSION

The experimental measurements, collected over different rain regimes in Brazil, were very useful in classifying and regionally defining the cloud processes of warm, deep and stratiform clouds. The cloud liquid water content of non-precipitating events in Northeast Brazil, in the coastal region, and in the regions characterized by significant subsidence, tends to generate higher values than those over the continent and in which deep convective processes are more noticeable. The adiabatic liquid water content obtained by radiosonde was very useful for inferring the importance of the entrainment and coalescence processes. In comparing the adiabatic liquid water estimation with that obtained by radiometry, we observed that clouds on coastal sites contain more measured liquid water than the clouds of continental sites, when compared according to adiabatic calculations. This result can be related to the occurrence of more important coalescence processes at the coastal sites, in contrast to more entrainment at the continental sites (less liquid water than the adiabatic calculations). This calculation is performed for non- raining clouds, where the large amounts of highly hygroscopic aerosols and the larger deep convective inhibition can favor this behavior. However, for deep convective systems, the continental clouds are significantly more efficient in the production of rainfall than the coastal clouds. The ILWC and thickness of the cloud layer have a positive relationship; however, the larger variations of ILWC with cloud thickness do not allow a single relationship to be established.

The integrated rain liquid water content (from the cloud base up to 1 km above the melting layer) was similar between Fortaleza and Alcantara, as well as between Vale do Paraiba and Belem. This reflects the individual characteristics of the precipitation regimes of the clouds in these regions. The most significant values were observed in the coastal region, which according to our analysis could be associated with warm cloud processes. Meanwhile, the upper layers showed higher values in Vale do Paraiba and Belem. This indicates the presence of cold cloud processes because these regions are known for their well-developed convective systems and high rain rates. With respect to the different types of rainy systems, it was noted that the LWCR and ILWR of the liquid water cloud layers were higher in convective events than other systems. Meanwhile, a small difference was observed among the ILWR for the warm and stratiform clouds. However, our regional analyses showed that the values of liquid water content were higher in the warm clouds for sites near the coast than those observed over the continental sites, particularly for Alcantara. The warm clouds in Vale had the lowest values of liquid water. Coastal stratiform events have high water content in the liquid water layer, most likely due to the presence of clouds with ice caps, causing them to be classified as stratiform but with warm cloud processes. It was noted that the clouds at Vale do Paraiba exhibit more intense processes in the mixed layer than any other cloud types. This is most likely due to the large amount of supercooled water; additionally, in this field campaign, several thunderstorm and hailstorms were reported. In Belem, we found clouds with more important glaciated layers and high cloud tops. The stratiform clouds showed different brightness band signatures. The clouds in the coastal regions presented closer brightness bands to the melting layer than the others. This is most likely due to the occurrence of faster ice melting here than in the regions where deep convective clouds are more pronounced.

With respect to raindrop size, it was noted that, except for stratiform clouds, convective events, whether shallow or deep, vary significantly between sites. Such behavior reflects the dominant microphysical processes of each region. The DSDs for warm clouds on the coast confirm that higher ILWs are associated with larger raindrops. A specific range of raindrops can cause differences in the rainfall observed at each site, as well as the LWC values. It was noted that certain diameters have a greater effect on the total liquid water content, which depends on the system observed and the location where the rain was generated. This is a reflection of the predominant rain process in the cloud. The evaluation of the rain liquid water behavior in relation to the rain rate intensity showed that there is significant variability in the values associated with cloud type. It was also observed that as the rain intensity increases, the rain liquid water content is increased. The relationship between ILW and RR is strongly associated with the observed cloud type. However, developing a universal association with these variables is complex and requires ancillary information.

Although there are similarities among the mean water distributions of the studied sites, it was evident that significant variability exists in the cloud liquid water values and precipitation, which includes the microphysical, adiabatic and non-adiabatic processes at each site.

Acknowledgements. This work was supported by FAPESP grant No. 2009/15235-8 and CNPQ No. 140818/2011-1. We also thank the CHUVA campaign team for their efforts to help with equipment function.

6. REFERENCES

Atlas, D., 1954: The Estimation of Cloud Parameters by radar. J. Meteor. (11), 309– 317. Awaka, J., Iguchi, T., Okamoto, K., 2007. Rain Type Classification Algorithm, In Measuring Precipitation from Space. Advances in Global Change Research (28), pp 213-224. Battan, L.J., 1973: Radar Observation of the Atmosphere. The University of Chicago Press, Chicago. Caracciolo, C., Prodi, F., Uijlenhoet, R., 2006. Comparison between Pludix and impact/optical disdrometers during rainfall measurement campaigns, Atmospheric Research (82), Issues 1–2, 137-163. Cimini, D., Westwater, E.R., Han, Y., Keihm, S.J., 2003. Accuracy of Ground-based Microwave Radiometer and Balloon-Borne Measurements During WVIOP2000 Field Experiment, IEEE T. Geosci. Remote (41), 2605–2615. Czekala, H. Simmer, C., 1998. Microwave radiative transfer with nonspherical precipitating hydrometeors, J. Quant. Spectrosc. Radiat. Transfer, 60(3), 365–374. Czekala, H., Crewell, S., Simmer, C., Thiele, A., 2001. Discrimination of cloud and rain liquid water path by groundbased polarized microwave radiometry, Geophys. Res. Lett., 28(2), 267–270. Cotton, W.R., Bryan, G.H., van den Heever, S.C., 2011. Storm and cloud dynamics, Academic Press. Crewell, S., Loehnert, U., 2003. Accuracy of cloud liquid water path from ground-based microwave radiometry: 2. Sensor accuracy and synergy, Radio Sci., 38(3), 8042. Donaldson, R.J., 1955. The Measurement of Cloud Liquid-Water Content by radar. J. Meteor., 12, 238–244. Doviak, R.J., Zrnic, D.S., 1993. Doppler Radar and Weather Observations. Academic Press New York, NY, USA, 2nd edition, 562 pp. Ebell, K., Löhnert, U., Crewell, S., Turner, D.D., 2010. On characterizing the error in a remotely sensed liquid water content profile, Atmospheric Research (98), Issue 1, 57- 68. Eccles, P.J., Mueller, E.A., 1971. X-Band Attenuation and Liquid Water Content Estimation by a Dual-Wavelength Radar. J. Appl. Meteor. (10), 1252–1259. Fabry, F., Zawadzki, I., 1995. Long-Term Radar Observations of the Melting Layer of Precipitation and Their Interpretation. J. Atmos. Sci. (52), 838–851. Greene, R., Robert, D.R., Clark, A., 1972. Vertically Integrated Liquid Water – A New Analysis Tool. Mon. Wea. Rev. (100), 548–552. Hagen, M., Yuter, S.E., 2003. Relations between radar reflectivity, liquid-water content, and rainfall rate during the MAP SOP. Q.J.R. Meteorol. Soc. (129), 477–493. doi: 10.1256/qj.02.23 Han, Y., Westwater, E.R., 2000. Analysis and improvement of tipping calibration for ground-based microwave radiometers. Geoscience and Remote Sensing, IEEE Tras. (38), Issue 3, 1260 – 1276. Hardy, W.N., 1973. Precision Temperature Reference for Microwave Radiometry. IEEE Trans. Microwave Theory (21), 149-150. Hewison, T.J., 2006. Profiling Temperature and Humidity by Ground-based Microwave Radiometers, PhD Thesis, University of Reading. Hogan, R.J., Gaussiat, N., Illingworth, A.J., 2005. Stratocumulus Liquid Water Content from Dual-Wavelength Radar. J. Atmos. Oceanic Technol. (22), 1207–1218. Houze, R. A. Jr., 1993. Cloud Dynamics. 573 pp., Academic, San Diego. Hu, Z., Srivastava, R.C., 1995. Evolution of raindrop size distribution by coalescence, breakup, and evaporation: theory and observations. Journal of the Atmospheric Sciences (52), 1761–1783. Ingold, T., Peter, R., Kämpfer, N., 1998. Weighted mean tropospheric temperature and transmittance determination at millimeter-wave frequencies for ground-based applications, Radio Sci. (33), 905-918. Islam, T., Rico-Ramirez, M.A., Thurai, M., Han, D., 2012. Characteristics of raindrop spectra as normalized gamma distribution from a Joss–Waldvogel disdrometer, Atmospheric Research (108), 57-73. Karmakar, P.K., Maiti, M., Sett., S., Angelis, C.F., Machado, L.A.T., 2011. Radiometric estimation of water vapor content over Brazil. J. Adv. Space Res. (28), 1506-1514. Korolev, A. V., Isaac, G. A., Strapp, J. W., Cober, S. G., Barker, H. W., 2007. In situ measurements of liquid water content profiles in midlatitude stratiform clouds. Q.J.R. Meteorol. Soc. (133), 1693–1699. doi: 10.1002/qj.147. Lawson, R.P., Blyth, A.M., 1998. A comparison of optical measurements of liquid water content and drop size distribution in adiabatic regions of Florida cumuli, Atmospheric Research (47–48), 671-690. Leuenberger, A., 2009. Precipitation Measurements with Microwave Sensors. Bern University. Liljegren, J., Clothiaux, E., Mace, G., Kato, S., Dong, X., 2001. A new retrieval for cloud liquid water path using a groundbased microwave radiometer and measurements of cloud temperature, J. Geophys. Res. (106), 14485–14500. Löhnert, U., Crewell, S., Simmer, C., Macke, A., 2001. Profiling Cloud Liquid Water by Combining Active and Passive Microwave Measurements with Cloud Model Statistics. J. Atmos. Oceanic Technol. (18), 1354–1366. Löhnert, U., Crewell, S., 2003. Accuracy of cloud liquid water path from ground-based microwave radiometry 1. Dependency on cloud model statistics, Radio Sci., 38, 8041. Marshall, J.S.; Gunn, K.L.S., 1952. Measurement of snow parameters by radar. Journal of the Atmospheric Sciences (9), 5, 322–327. Marshall, J.S., Palmer, W.McK., 1948. The distribution of raindrops with size. J. Meteor. (5), 165–166. Martins, C.R., Machado, L.A.T., Costa, A.A., 2010. Characterization of the microphysics of precipitation over Amazon region using radar and disdrometer data. (96), Issues 2-3, 388-394. Matzler, C., Morland, J., 2009. Refined Physical Retrieval of Integrated Water Vapor and Cloud Liquid for Microwave Radiometer Data. IEEE Transactions on Geoscience and Remote Sensing 47:6, 1585-1594. Meywerk, J.,Quante, M., Sievers, O., 2005. Radar based remote sensing of cloud liquid water—application of various techniques—a case study, Atmospheric Research (75), Issue 3, 167-181. Michaelides, S., Levizzani, V., Anagnostou, E., Bauer, P., Kasparis, T., Lane, J.E., 2009. Precipitation: Measurement, remote sensing, climatology and modeling, Atmospheric Research (94), Issue 4, 512-533. Peter, R., Kämpfer, N., 1992. Radiometric determination of water vapor and liquid water and its validation with other techniques, J. Geophys. Res., Vol. 97, No. D16, 18173-18183. Peters, G., Fischer, B., Münster, H., Clemens, M., Wagner, A., 2005. Profiles of Raindrop Size Distributions as Retrieved by Microrain Radars. J. Appl. Meteor. (44), 1930-1949. Peters, G., Fischer, B., Clemens, M., 2010. Rain Attenuation of Radar Echoes Considering Finite-Range Resolution and Using Drop Size Distributions. Journal of Atmospheric and Oceanic Technology 27:5, 829-842. Pruppacher, H.R., Klett, J.D., 1998. Microphysics of Clouds and Precipitation. 2nd Edn. Kluwer Academic Publishers, Boston. Rogers, R.R., Yau, M.K., 1989. A Short Course in Cloud Physics. Pergamon Press, 293 pp. Rose, T., Crewell, S., Löhnert, U., and Simmer, C., 2005. A network suitable microwave radiometer for operational monitoring of the cloudy atmosphere, Atmos. Res. (75), 183-200. Rossow, W.B., Schiffer, R.A., 1999. Advances in understanding clouds from ISCCP, Bull. Amer. Meteor. Soc. (80), 2261-2287. Saavedra, P., Battaglia, A., Simmer, C., 2012. Partitioning of cloud water and rainwater content by ground-based observations with the Advanced Microwave Radiometer for Rain Identification (ADMIRARI) in synergy with a micro rain radar, J. Geophys. Res. (117), 1984-2012. Schneebeli, M., Sakuragi, J., Biscaro, T., Angelis, C. F., Carvalho da Costa, I., Morales, C., Baldini, L., Machado, L. A. T., 2012. Polarimetric X-band weather radar measurements in the tropics: radome and rain attenuation correction, Atmos. Meas. Tech. (5), 2183-2199. Solheim, F., Godwin, J., Westwater, E., Han, Y., Keihm, S., Marsh, K., Ware, R., 1998. Radiometric profiling of temperature, water vapor, and liquid water using various inversion methods. Radio Sci. (33), 393–404. Skou, N., Vine, D., 2006. Microwave radiometer systems: design and analysis. 2nd ed. ISBN 1-58053-974-2. Tokay, A., Short, D.A., 1996. Evidence from Tropical Raindrop Spectra of the Origin of Rain from Stratiform versus Convective Clouds. J. Appl. Meteor. (35), 355–371. Tokay, A., Kruger, A., Krajewski, W.F., 2001. Comparison of Drop Size Distribution Measurements by Impact and Optical Disdrometers. J. Appl. Meteor. (40), 2083–2097. Ware, R., Carpenter, R., Guldner, J., Liljegren, J., Nehrkorn, T., Solheim, F., Vandenberghe, F., 2003. A multichannel radiometric profiler of temperature, humidity, and cloud liquid. Radio Sci. (38), 8079, doi:10.1029/2002RS002856. Westwater, E., 1993. Ground-based microwave remote sensing of meteorological variables, In Atmospheric Remote Sensing by Microwave Radiometry, edited by M. Janssen, chap. 4, 145–213, John Wiley, New York. Westwater, E., Crewell, S., Matzler, C., Cimini, D., 2005. Principles of surface-based microwave and millimeter wave radiometric remote sensing of the troposphere. Quat. Soc. Ital. Elettromagn. (1), 50-90. 8041, doi:10.1029/2002RS002654. Won, H.Y., Kim, Y-H., Lee, H.S., 2009. An application od brightness temperature received from a ground-based microwave radiometer to estimation of precipitation occurrence. Asian-Pacific Journal of atmospheric (45), 1, 55-69. Zhao, G., Chu, R., Zhang, T., Jia, W., 2013. Rainwater content estimated using polarimetric radar parameters in the Heihe River Basin, Atmospheric Research (120– 121), 155-161. Zhao, L., Weng, F., 2002. Retrieval of Ice Cloud Parameters Using the Advanced Microwave Sounding Unit. J. Appl. Meteor. (41), 384–395. Zhao, Q., Carr, F.H., 1997. A Prognostic Cloud Scheme for Operational NWP Models. Mon. Wea. Rev. (125), 1931-1953. Zrnic, D.S., Raghavan, R., Chandrasekar, V., 1994. Observations of Copolar Correlation Coefficient through a Bright Band at Vertical Incidence. J. Appl. Meteor., (33), 45-52.

Table and Figure Caption

Table 1. The cloud and rain liquid water contents by active and passive sensors for different regions and rainy systems in Brazil during the CHUVA campaign.

Table 2. The classifications* of rainy event occurrence (RR>0.1 mm/h) over the main site for all CHUVA Project experiments.

Figure 1. The sites locations and a schematic representation of the measurement strategy for the CHUVA campaign from 2010 to 2012.

Figure 2. The relative frequency histogram of the cloud liquid water content for all sites during the CHUVA campaign, estimated by (a) microwave ground-based radiometer

MP3000A (ILWC, mm) and (b) the adiabatic approach from radiosondes (ILWadia, mm).

Figure 3. The ILWC (mm) estimated by the radiometer, related to the cloud thickness (m) retrieved by LiDAR in the non-precipitation condition at Fortaleza.

Figure 4. The relative frequency histogram of the rain liquid water content (ILWR, mm) integrated up to 1 km below the 0°C height for all sites during the CHUVA campaign, estimated by (a) MRR and (b) XPOL.

Figure 5. The relative frequency histogram of the rain liquid water content (ILWR, mm) integrated up to 1 km below the 0°C height for different rainy systems, estimated by XPOL for each site for (a) warm rain, (b) stratiform, and (c) convective.

Figure 6. The mean profile of the rain liquid water content (LWCR, mm) below the 0°C height for different rainy systems, estimated by XPOL for each site for (a) warm rain, (b) stratiform, and (c) convective. Figure 7. (a-c) The mean raindrop concentrations (mm-6m-3) and (d-f) the relations (%) between the liquid water content for each diameter (Di) and the total observed for each DSD from the disdrometer measurements for warm (a,d), stratiform (b,e), and convective (c,f) clouds for all sites during the CHUVA project.

Figure 8. The integrated rain liquid water (ILWR, mm) estimated by XPOL for different rain intensities, including light rain (1 to 2.5 mm/h), moderate rain (2.5 to 10 mm/h), heavy rain (10 to 50 mm/h), and severe rain (larger than 50 mm/h) for all sites in the CHUVA campaign. Figure and Tavles

Figure 1. The sites locations and a schematic representation of the measurement strategy for the CHUVA campaign from 2010 to 2012.

Figure 2. The relative frequency histogram of the cloud liquid water content for all sites during the CHUVA campaign, estimated by (a) microwave ground-based radiometer

MP3000A (ILWC, mm) and (b) the adiabatic approach from radiosondes (ILWadia, mm).

Table 1. The cloud and rain liquid water contents by active and passive sensors for different regions and rainy systems in Brazil during the CHUVA campaign. Integrated Liquid Water (mm) Non rainy Rainy MRR XPOL Site ILW ILW C adia ILW (H –H ) VIL MP3000A Radiosonde R LCL 0°C-1km General Warm Stratiform (BB*) Deep Convection Mixed Glaciated Mean 0.23 0.19 0.24 0.58 0.32 0.26 4.10 0.09 0.02 Fortaleza/CE Std 0.24 0.44 0.41 1.43 0.46 0.22 3.15 0.17 0.03 Mean 0.22 0.29 0.29 0.44 0.15 0.27 2.39 0.12 0.09 Belem/PA Std 0.25 0.41 0.44 0.80 0.19 0.15 1.34 0.29 0.17 Mean 0.36 0.22 - 0.47 0.30 0.43 2.03 0.06 0.01 Alcântara/MA Std 0.28 0.36 - 0.73 0.39 0.29 1.89 0.08 0.03 Mean 0.14 0.28 0.39 0.38 0.09 0.26 2.45 0.14 0.02 Vale do Paraíba/SP Std 0.15 0.50 1.26 0.83 0.11 0.19 1.84 0.54 0.06 Mean 0.25 Warm Rain Std 0.37 Stratiform (with Mean 0.29 BB) Std 0.22 Mean 3.02 Convective Std 2.52 BB – Brightness Band

Table 2. The classifications* of rainy event occurrence (RR>0.1 mm/h) over the main site for all CHUVA Project experiments.

Site Fortaleza Belem Vale Alcantara Type % RRMean % RRMean % RRMean % RRMean Stratiform (with BB) 36 1.8 19 1.8 27 2.4 26 3.7 Convective 8 46.2 8 61.6 6 62.5 6 27.5 Warm Pure 12 3.6 25 4.9 14 1.94 19 7.2 *The statistic is only related to these 3 distinct rainy classes; other events can be identified but are not computed here, such as stratiform without a bright band, weak convective, and warm with other clouds in the upper levels.

Figure 3. The ILWC (mm) estimated by the radiometer, related to the cloud thickness (m) retrieved by LiDAR in the non-precipitation condition at Fortaleza.

Figure 7. (a-c) The mean raindrop concentrations (mm-6m-3) and (d-f) the relations (%) between the liquid water content for each diameter (Di) and the total observed for each DSD from the disdrometer measurements for warm (a,d), stratiform (b,e), and convective (c,f) clouds for all sites during the CHUVA project.