**5. Cloud studies**

Clouds play an important role in Earth's radiation budget as they reflect incoming solar radiation and absorb outgoing thermal radiation emitted from the surface, the atmosphere, and other clouds [71]. In tropical regions, Cirrus clouds are omnipresent [72] and hence affect the radiation balance significantly because of their sizable horizontal extent (hundreds to thousands of km) and long lifetime (hours to days). Cirrus optical properties, altitude, vertical and horizontal coverage control, radiative forcing, and detailed measurements are of absolute value [73, 74].

Studies reporting cirrus properties over tropical South America used to be scarce, with most studies based on in-frequent satellite observations [72, 75]. For obtaining detailed geometrical and microphysical properties of Cirrus clouds,

especially the sub-visible type, ground-based lidars are indispensable. For this reason, many lidar groups around the world have used such measurements to obtain and report these characteristics in mid-latitudes [74] and tropical regions [76]. A similar effort is being made in South America, mostly under the auspices of LALINET. Here we review studies from five LALINET stations (Cuba, Manaus, Natal, São Paulo, and Punta Arenas), and we also discuss results from Reference [77], who report on measurements in Buenos Aires. It is compiled a set of statistics for optical and geometrical properties of cirrus occurring in this side of the world, summarized in **Table 2**.

The first studies were performed in Cuba by the lidar group in Camagüey using 6-years (1993 to 1998) elastic lidar data [78]. The lidar measurements were performed for detecting aerosols in the stratosphere, and hence were conducted mostly on clear nights to the naked eye, thus introducing a bias in the cloud measurements towards subvisual cirrus clouds. Indeed, from 131 clouds measured, only 8% were thick (COD >0.3), 67% were thin (0.03 < COD <0.3) and 25% were sub-visual (COD <0.03). Sub-visual and thin cirrus have an average cloud top and base at 14.1 and 11.6 km, respectively, with thick clouds occurring at slightly lower altitudes. The authors estimated the respective optical depths to be 0.50 ± 0.27, 0.07 ± 0.05, and 0.02 ± 0.01, but these are an upper limit as they were calculated assuming a lidar ratio (LR) of 10 sr, which we now know is too low.

In a follow-up study, Barja and Antuña performed radiative transfer simulations (GFDL model, Freidenreich and Ramaswamy, 1999) of the impact of cirrus clouds on solar radiation (SW) [83]. They have found that the daily mean value of SW cirrus radiative forcing (CRFSW) has an average value of −9.1 W m−2 at the top of the atmosphere (TOA) and −5.6 W m−2 at the surface (SFC). There is a linear relation between CRFSW and COD, with a slope of - 30 W m−2/COD at TOA. The local radiative heating effect where the cirrus is found ranged from 0.35 to 1.24 K day−1, with an average of 0.63 K day−1. These results were the first to show that tropical cirrus radiative impact on Earth's energy balance is essential.

In the Amazon region, Gouveia et al. used continuous measurements (July 2011 to June 2012) at the Manaus station [1] to retrieve the optical and geometrical properties of cirrus clouds during day and night [79]. The cirrus frequency of occurrence was about 88% during the wet season and not lower than 50% during the dry season, with a mean column optical depth of 0.25 ± 0.46. The cloud top and base heights, as well as cloud thickness and cloud optical depth, were, respectively, 14.3 ± 1.9 (std) km, 12.9 ± 2.2 km, 1.4 ± 1.1 km, and 0.25 ± 0.46, similar to the values reported in Cuba. These clouds have a significant radiative impact with such a high frequency of occurrence and altitude over the dark-pristine Amazon forest (albedo about 0.12).

Gouveia then used these measured optical depths and vertical profiles of the cirrus extinction coefficient [84], at 5-min and 200 m resolutions, and calculated the cirrus radiative forcing (CRF) with libRadtran [85]. Cirrus in the Amazon region produced a net CRF at TOA and SFC of +15.3 and − 3.7 W m−2, respectively, much more intense than predicted for 3 European sites (+0.9, +1, and + 1.7 W m−2 at TOA) [74, 86]. Optically thicker cirrus, in general, have more prominent net CRF, with instantaneous CRF that could reach extremes up (down) to +140 (−65) W m−2 for the night (day) time [85]. Together, the vertical profiles with total COD >0.3 were responsible for about 72% (62%) of the TOA (BOA) net CRF, which means that a large fraction of the CRF is generated by optically thin cirrus (COD <0.3) that are harder to detect by radars and passive instruments on board of satellites [86]. A definite daily cycle of the optical depth was found and shows a minimum about local noon and a maximum in the late afternoon (~16 h LT), associated with the diurnal precipitation cycle. This results in a mean instantaneous TOA (SFC)


*Lidar Observations in South America. Part II - Troposphere DOI: http://dx.doi.org/10.5772/intechopen.95451*

*Cirrus clouds' properties were measured by different LALINET stations, from north to south, during the last 15 years. Columns indicate latitude (deg) of the station, period of study and wavelength, frequency of occurrence, cirrus layer average top and base altitudes, column cloud optical depth (COD), and lidar ratio (LR).* net CRF ranging from +1.7 (−23) W m−2 in the afternoon to +47 (+3.1) W m−2 at night [86]. The cirrus clouds produced an average in-cloud heating rate ranging between −1 K day−1 to +2 K day−1 vertical profile from 8 to 18 km (in-cloud), but with instantaneous values that can reach values higher than 10 K day−1 for portions of the cloud with high ice water content [86].

The DUSTER Lidar station at Natal, Brazil, is the most recent addition to the network. Cirrus measurements started only recently, and Santos reports on measurements during January–February (pre-rainy season) of 2017 and 2018 [80]. A total number of 35 clouds were studied and showed cirrus up to 16 km. These clouds are lower than those in Manaus and were never observed above the tropopause at 17–18 km. The reason is the less vigorous convection in this coastal site, which resembles an oceanic precipitation regime. The frequency of occurrence and average cloud top altitude was 74% (57%) and 13.9 km (12.3 km) in 2017 (2018), respectively. In situ data obtained by radiosondes (9.5 km away) for selected case studies showed an increase of the relative humidity in the layer where the lidar identified the cirrus clouds, from around 10–20% below/above the cloud to around 40–55% in the cloud altitude region.

Cirrus clouds over São Paulo, in the subtropics of South America, were studied by Larroza [81]. She analyzed 34 days, from June to July 2007, using the methodology described in Ref. [87]. The cirrus frequency of occurrence was 54%. The vertical distribution of cloud tops showed peaks at 9.6, 10.6, 12.3, and 13.9 km, with an average of 12.4 km. These cirri were optically thinner (0.27) and occurred at lower altitudes (cloud top 12.4 km) than their tropical counterpart but had a similar lidar ratio of about 26 sr. The clouds observed were either produced by the passage of cold fronts or transported from the tropics or mid-latitudes.

Going into the mid-latitudes, Lakkis et al. used data from a lidar system in Buenos Aires, Argentina, that was not part of LALINET [77]. They studied 60 diurnal cirrus clouds from 2001 to 2005. Unlike the tropics and sub-tropics, cirrus tops were only found very close to the tropopause (~380 m), with cloud tops at 11.8 ± 0.86 and bases at 9.6 ± 0.9 km. Unfortunately, the low statistics did not allow the calculation of a frequency of occurrence, nor did the authors report values of optical depth or lidar ratios. The southernmost LALINET station is at Punta Arenas, Chile (53°S, 71°W), a sub-Antarctic region. Lidar cirrus measurements there began in October 2016 and continue to the present. A preliminary result of cirrus clouds' geometric characteristics in the region, over two years (from 2016 to 2018), shows that the cirrus' mean base height is 9.0 ± 2.4 km and the mean top height is 10.8 ± 2.2 km. In the same site in November 2018, the Leipzig Aerosol and Cloud Remote Observations System (LACROS) [88] was deployed by the Leibniz Institute for Tropospheric Research (TROPOS) in collaboration with the University of Magallanes (UMAG) for the field experiment DACAPO - PESO (Dynamics, Aerosol, Cloud And Precipitation Observations in the Pristine Environment of the Southern Ocean). Cirrus clouds measurements are being performed with a Raman polarization lidar that will allow the calculation of LR and COD at multiple wavelengths, which will be reported in a future study.

There has been a great effort to study cirrus clouds in South America, with measurements taking place from 1993 to today, and from 21 N to 53 S. **Table 3** summarizes the main characteristics, and we can see how the cirrus altitude changes from the Tropical (high, 14.3 km) to the sub-Antarctic regions (low 10.8 km). The frequency of occurrence also becomes smaller, reducing from 74% in Manaus to 45% in Punta Arenas. Deep convection, prevalent in the tropics and sub-tropics, pushes the tropopause upward and creates optically and physically thick cirrus clouds from the anvil's detrainment. From the subtropics towards the polar regions, convection is limited by the lack of surface heating, and the mixing of tropospheric air depends on


*Lidar Observations in South America. Part II - Troposphere DOI: http://dx.doi.org/10.5772/intechopen.95451*

**Table 3.**

*Cirrus Clouds radiative forcing (W m−2) over Cuba [83] and the central Amazon forest [79] calculated by radiative transfer simulations based on ground-based lidar measurements.*

uplift by frontal systems. This does not reach high altitudes, and the primary cirrus production mechanism will be the large-scale lifting of water vapor, rendering physically and optically thinner clouds. Lidars can directly observe the optical depth, and **Table 3** shows that it also becomes smaller, decreasing from 0.35 in Manaus to 0.25 in Punta Arenas. Unfortunately, there are not enough LR estimations to allow for a comparison. They were calculated only for Manaus and São Paulo, and the results are in close agreement, indicating similar crystal habits and formation mechanisms, as expected. Similar relation of the cirrus features and latitude was first reported by Cordoba-Jabonero et al. using data from the LALINET subtropical station of São Paulo and ground-based lidar located in Belgrano Antarctic Station [88].

Another aspect of this LALINET effort is the diversity in the methods and the timespan of the different studies, limiting our ability for a more in-depth comparison. The combination of the Klett and transmittance methods, as described by Larroza [87] and Gouveia et al. [79], can be applied to elastic data from any of the LALINET lidars, providing COD and LR with high temporal resolution (e.g., 5-min), during both day and night. The use of radiative transfer models to calculate the cirrus' radiative impact could also be performed for all stations doing cirrus measurements. It would be essential to homogenize cirrus clouds' analysis throughout the network groups, even by sharing analysis algorithms. Moreover, it should be emphasized that these critical analyses could be automatized and performed unattended on a unique central server. These facts highlight the importance of establishing systematic data sharing in the context of LALINET and GALION.
