Mobile Lidar Profiling of Tropical Aerosols and Clouds - American ...

1288

JOURNAL OF ATMOSPHERIC AND OCEANIC TECHNOLOGY

VOLUME 25

Mobile Lidar Profiling of Tropical Aerosols and Clouds P. C. S. DEVARA, P. E. RAJ, K. K. DANI, G. PANDITHURAI, M. C. R. KALAPUREDDY, S. M. SONBAWNE, Y. J. RAO, AND S. K. SAHA Indian Institute of Tropical Meteorology, Pashan, Pune, India (Manuscript received 21 February 2007, in final form 19 June 2007) ABSTRACT Lidar profiling of atmospheric aerosols and clouds in the lower atmosphere has been in progress at the Indian Institute of Tropical Meteorology (IITM), Pune (18°32⬘N, 73°52⬘E, 559 m MSL), India, for more than two decades. To enlarge the scope of these studies, an eye-safe new portable dual polarization micropulse lidar (DPMPL) has been developed and installed at this station. The system utilizes a diodepumped solid-state (DPSS) neodymium–yttrium–aluminum–garnet (Nd:YAG) laser second harmonic, with either parallel polarization or alternate parallel and perpendicular polarization, as a transmitter and a Schmidt–Cassegrain telescope, with a high-speed detection and data acquisition and processing system, as a receiver. This online system in real-time mode provides backscatter intensity profiles up to about 75 km at every minute in both parallel and perpendicular polarization channels, corresponding to each state of polarization of the transmitted laser radiation. Thus, this versatile lidar system is expected to play a vital role not only in atmospheric aerosol and cloud physics research and environmental monitoring but also in weather and climate modeling studies of the impact of radiative forcing on the earth–atmosphere radiation balance and hydrological cycle. This paper provides a detailed description of Asia’s only lidar facility and presents initial observations of space–time variations of boundary layer structure from experiments carried out during winter 2005/06.

1. Introduction Aerosols of both natural and anthropogenic origin perturb the atmospheric radiation field through direct and indirect interactions with solar radiation (Charlson and Heintzenberg 1995; Ramanathan et al. 1989). Monitoring of the impacts of natural aerosols can also help in understanding the evolution of past environments and predicting future climate. Moreover, atmospheric aerosol characteristics vary significantly in space and time over different environments. Thus, accounting for the effects of aerosols on the earth– atmosphere radiation balance and on environmental pollution and air quality assessment is a very complex and challenging exercise. For several reasons, the study of aerosols and clouds has special significance over the tropics, where convective and high-altitude dynamical processes influence the distribution of aerosols. The

Corresponding author address: Dr. P. C. S. Devara, Indian Institute of Tropical Meteorology, Dr. Homi Bhabha Road, NCL Post Office, Pashan, Pune 411 008, India. E-mail: [email protected] DOI: 10.1175/2007JTECHA995.1 © 2008 American Meteorological Society

aerosols that are lifted because of daytime convective activity are suspended for a considerable amount of time in the lowest layers of the atmosphere, and those generated near the surface return to lower levels during the nighttime. During the night, the aerosol layers are horizontally stratified because mixing is limited by the surface radiation inversions. The nighttime radiative cooling of the surface and of the air aloft induces stable stratification near the ground, and advection and/or subsidence begins to play a major role in determining the aerosol concentration aloft (Mahrt 1985). Thus, the thermodynamical forcing influences the aerosol patterns that are formed because of surface-generated aerosols, especially during the early morning transition from a stable to convective boundary layer and the late evening transition from a convective to a stable boundary layer (Lenshow et al. 1979). Lidars play an important role in these studies because of their capability to make very precise continuous measurements of different aerosol and cloud parameters (McCormick et al. 1993). Detailed knowledge of aerosols and clouds is necessary mainly for obtaining better radiative forcing estimates—one of the major uncertainties in under-

AUGUST 2008

DEVARA ET AL.

1289

FIG. 1. Optical layout of the DPMPL at IITM, Pune, India.

standing the influence of aerosols and precursor gases on weather, climate change, and underlying processes—and for refining models for improving satellite data retrieval algorithms. In view of the importance of aerosols in tropical atmospheric processes (Hansen et al. 2000), the availability of data describing their main properties is rather poor, in particular with respect to their vertical distributions. Taking into account these requirements, among others, a bistatic Argon ion lidar system has been developed, and vertical profile measurements of aerosol number density have been made at the Indian Institute of Tropical Meteorology (IITM), Pune, India continually since 1985. Utilizing more than 1500 vertical profiles of lidar-observed aerosol concentration archived during October 1986–September 2000, a tropospheric aerosol climatology has been established (Devara et al. 2002). Using this multiyear lidar aerosol data, interannual, intraseasonal, and long-term trends in aerosol loading, the aerosol–cloud–precipitation relationship, and the air quality over the experimental station have all been investigated (Devara et al. 2003). Although a considerable number of studies have

been done on the effects on weather and climate of direct radiative forcing due to aerosols, studies of aerosol-related semidirect and indirect radiative forcing are very sparse (Houghton et al. 2001). Moreover, the most important component that is missing so far in tropical aerosol research, particularly in India, is multidimensional mapping of aerosol properties and cloud structures during both day and night over different environments (associated with complex terrain and meteorological conditions). In this context, the dual polarization micropulse lidar (DPMPL) system at IITM plays a vital role in atmospheric aerosol and cloud physics research (especially cirrus cloud characterization) and environmental monitoring. It also provides very valuable input information to weather, climate, and air quality models (Holm 2004; Kamineni et al. 2003; Beninston et al. 1990), especially those aimed at accounting for radiative forcing and its impact on the hydrological cycle on different spatial and temporal scales. A detailed description of the lidar, together with some first results, is presented in this paper. To the best of our knowledge, this DPMPL is the first of its kind available in Asia.

1290


VOLUME 25

2. System description and capabilities The system was built by following a uni-axial monostatic configuration. Because one objective was to conduct field campaigns at multiple sites, the system was designed to be eye-safe and mobile. This was achieved using a low-energy, high-repetition-rate Nd:YAG laser with an expanded beam. The receiver is a compact Schmidt–Cassegrain telescope with a focal ratio of f/10. Figure 1 shows the optical layout of the system. The entire system is composed of two basic parts. The first part includes the transmitter, receiver, electrical supply, and electronics, and is mounted on a vibration-isolated platform on casters in a thermoelectrically cooled and clean environment; the second is a high-reliability transportable control and data processing system. All of the hardware is controlled via software under the Microsoft Windows XP environment, and the majority of the controls, especially the high-speed (⬃500 MHz) acquisition sequences, are fully automated. The software utilizes a simple user-friendly graphics interface that makes the system easy to operate. The transmitter–receiver axis alignment is achieved by means of an “octopus,” which is a custom-made, highperformance, microcontroller-based remote terminal unit that essentially controls the x and y axes of a mirror to align the laser beam with the receiver. It communicates with the main computer via RS232 serial port and liaisons with photomultiplier tube (PMT) detectors, polarization rotators, alignment systems, and Fabry–Perot stepper motors (depending on whether lidar operation is required during daytime or nighttime or both) and executes their controls under instructions from the computer. The polarization rotator used in the present system was designed to meet our specific scientific goals. It flips the energy of each alternate laser pulse at a particular frequency between the parallel and perpendicular states of polarization by transmitting the laser beam through a Pockels cell and switching it by applying a high potential to the Pockels cell. To achieve this, a potential of about 5 kV was applied to the Pockels cell in the present system. A photograph depicting the complete transmitter–receiver and the interface for the data acquisition system of the DPMPL is presented in Fig. 2. The bore-sight mechanism of the system provides adjustment in two axes and maintains the alignment between the laser and receiver. (More details are presented in Table 1.) Moreover, the complete system can be tilted by a few degrees from the vertical before acquiring the data to avoid specular reflection, which might occur from horizontally oriented ice crystals during high-altitude cloud studies (e.g., Sassen 1991a). The system has built-in

FIG. 2. The outer view of the DPMPL facility and its different subunits: 1 ⫽ rollers and thermoelectric air conditioner; 2 ⫽ vibration isolation transmitter, receiver optics, and cabinet; 3 ⫽ laser controller and interface; 4 ⫽ on/off and emergency panel; 5 ⫽ alignment system; 6 ⫽ beam expander telescope; 7 ⫽ electrical and utility cabinet; 8 ⫽ 14-in. diameter telescope; 9 ⫽ power and computer interface; 0 ⫽ system tilt mechanism.

provision for applying corrections to the observed data resulting from the background and dark count. In the real-time (unattended) mode of operation, the system continuously acquires raw backscattered intensity (photon count) profiles for every minute in accordance with the prescribed altitude range and resolution settings. It is possible to select any altitude range of interest from the total vertical profile for detailed analysis. The finest range resolution that can be achieved with the system is 0.3 m (30 cm). Once the range resolution is set (depending on experimental requirements), it is maintained throughout the set altitude range. Thus, the data flow is very high during the high-resolution data recording periods.

3. Synthesis of data The signal detected by the lidar at each altitude is described in terms of the range-squared corrected signal (RSCS), which is usually defined as RSCS ⫽ 关P共␭, z兲 ⫺ P0共␭, z兲兴 ⫻ z2,

共1兲

where P(␭, z) and P0(␭, z) denote the power of the backscattered light and the background signal from an altitude of z, respectively. These profiles were also treated for air molecules produced by Rayleigh scattering with values taken from the standard atmosphere (Sasi and Sen Gupta 1986; Sasi 1994). A criterion of backscatter ratio in both of the co- and crosspolarization channels has been established to distin-

AUGUST 2008

1291

DEVARA ET AL. TABLE 1. Main specifications of DPMPL. Transmitter

Laser type: DPSS Nd:YAG Laser wavelength: 532 nm Laser repetition rate: 2 kHz (for 75 km) max. 50 kHz Pulse energy @ 10 kHz: 21 ␮J pulse⫺1 Pulse width at 2 kHz PRR: 18.3 ns Beam profile: TEM00 Polarization ratio and direction: 100:1; vertical Beam expansion: ⬎20 Polarization flipper: alternate parallel and perpendicular Polarization switching: 1 kHz at 2 kHz laser PRR Detectors Type: metal package PMTs Gain: 1.0 ⫻ 106 at 8 V (approx) Set point and monitor: Octopus Control by computer Data processing computer Speed: 42.8 GHz Memory: 1 GB

Receiver

Data acquisition

Optical design: telescope (Schmidt–Cassegrain) Clear aperture: 355.6 mm Focal length: 3556 mm

Method: photon counting Interface type: dual-channel MCS Time bin resolution (chosen in powers of 2): 2–8 ms Range resolution: 0.3 m Dynamic range: 35 bits at 2 ns Dead time at end of sweep: ⬍200 ns Dead time between time bins: none

Focal ratio: f/10 Transmitter receiver coupling: direct Filter bandwidth: 0.6 nm Fabry–Perot port: available Fabry–Perot driver: stepper motor available Fabry–Perot control: Octopus

Energy monitor Integrated with the system Interface: USB Software: stand alone

Alignment Mechanism: bore-sight Position sensing 4 quadrant PSD Alignment axis: x–y (stepper motors) Mechanism control: Octopus

Utilities Vibration isolation: 8⫻ isolation columns Cooling (optics and laser detectors): solid-state thermoelectric cooling (TEC) system

Hard disk: 110 GB Graphic card: NVIDIA GeForce MX 4000 Monitor: LCD 17⬙ Environment: Windows XP Professional SP2

guish clouds from the clear-air echoes. Also, the standard deviation (␴RSCS) is calculated from the temporal fluctuation of RSCS at each altitude to distinguish the lidar returns from clear air and clouds. These datasets are further subjected to the extinction coefficient (Fernald 1984; Klett 1981) and linear depolarization ratio (Schotland et al. 1971; Sassen 1991b) analyses. In general, even when the incident light wave is plane polarized with the plane of polarization parallel or perpendicular to the scattering plane, scattered radiation will contain both parallel and perpendicular polarized components. This is mainly because of the anisotropy of aerosol scattering. If the individual aerosols are assumed to be isotropic (spherical), the polarization components along the principal direction are equal, and the components in all other directions vanish. This implies that for isotropic particles, if the incident wave is plane polarized with the plane of polarization parallel to the scattering plane, the scattered radiation contains only the parallel component; and if the incident wave is plane polarized with the plane of polarization perpendicular to the scattering plane, the scattered radiation contains only the perpendicular component. This means that no depolarization occurs. Thus, the amount of depolarization is a measure of the anisotropy of the

scatterer. The degree of polarization gives the relative contribution of each polarization component resulting from the isotropy or anisotropy of the scatterer. Thus, the scattering properties of atmospheric aerosols differ significantly with the state of polarization of incident laser radiation. As explained above, for an incident nonpolarized light, the parallel and perpendicular polarization components will be affected unequally by the scattering phenomenon. This amounts to some polarization effect for the incident nonpolarized light. By making lidar measurements of aerosols with both parallel and perpendicular polarized laser light, it is possible to study parameters such as the degree of polarization and the depolarization ratio. These parameters are very useful for studying the isotropy or anisotropy of scattering characteristics of atmospheric aerosols during different weather conditions. These parameters also provide information on the microphysics of raining and/or nonraining clouds that is not limited to their mean phase (liquid or ice or mixed phase). The linear depolarization ratio, which is the calibrated ratio of signals in the perpendicular polarization plane to those in the parallel polarization plane, provides a semiquantitative indicator of particle morphol-

1292


VOLUME 25

FIG. 3. Sample profiles of backscatter intensity in both co- and cross-polarization channels of lidar (pulse width ⫽ 20 ns; pulse repetition rate (PRR) ⫽ 2000 Hz; time interval ⫽ 1 min; and pulse energy ⫽ 21 ␮J pulse⫺1 at 10 kHz) observed on a cloudy day.

ogy. From the lidar signals recorded in the P (parallel or copolarization) and S (perpendicular or crosspolarization) channels, these ratios at different altitudes (z) were computed by assuming a reasonable value of 0.028 (Bodhaine et al. 1999) for molecular depolarization correction for dry air and also by neglecting the aerosol contributions in the upper troposphere and stratosphere during nonvolcanic conditions. The backscatter ratio (RS) is computed as RS 共z兲 ⫽ ␤t 共z兲 Ⲑ␤m共z兲,

共2兲

where ␤t(z) is total backscatter containing contributions from both aerosols and air molecules. The backscatter coefficient is generally deduced from the measurements of backscatter light intensity (lidar), temperature, and pressure (local radiometersonde or

model atmosphere as explained above). The linear depolarization ratio (␦) is evaluated as

␦共z兲 ⫽ ␤⊥t 共z兲 Ⲑ 关␤㛳t 共z兲 ⫺ ␤㛳m共z兲兴,

共3兲

where the superscripts (⊥) and (㛳) denote the scatter coefficients estimated at perpendicular and parallel polarized planes to the polarization plane of the emitted laser pulse, respectively. To determine the cloud optical depth, the backscatter coefficients are vertically integrated over different cloud height range cells. Generally, the shape of aerosol particles is assumed to be spherical to simplify the calculation of certain scattering parameters like the differential Mie scattering cross section, single scattering albedo, etc. But in reality, apart from liquid aerosols, most aerosols, including dust particles over oceans, are nonspherical. To

AUGUST 2008

DEVARA ET AL.

1293

FIG. 4. Time evolution of raw backscattered energy observed in both cross- (perpendicular) and co- (parallel) polarization channels of the lidar on the night of 30/31 Dec 2005.

elucidate this aspect, Mishchenko et al. (1997) performed theoretical calculations and reported that more than 15% uncertainty in aerosol radiative forcing estimation might be caused by the above assumption. Hence, the shape of aerosol particles is an important parameter, and not many observations are available in the literature. In the present paper, we have made an attempt to infer the aerosol shape qualitatively from the lidar depolarization ratio. Also, by utilizing the unique facility (the switching of the state of polarization of the laser pulse energy between parallel and perpendicular) available with the DPMPL, datasets are being collected to undertake detailed analyses of cloud composition [such as determination of water, ice, or mixed phase and the shape and orientation of aerosol particles and extinction profiles (Sivakumar et al. 2003) as recorded with both co- and cross-polarization characteristics of the laser beam].

4. Sample results and discussion A typical profile of the lidar backscatter intensity acquired from surface to ⬃35 km altitude range with a

high-spatial-range resolution of 3 m is depicted in Fig. 3. It is interesting to see from this figure that in addition to the exponential decay (useful for extracting information on structure and stratification of atmospheric boundary layer) in the lidar return intensity in the lower-altitude region, a strong echo from a doublelayer cloud exists in both the P and S channels between 3 and 4 km. Such strong lidar echoes from clouds are thought to be caused by multiple scattering, possibly due to a larger cross-section of cloud droplets and to multiple internal reflections of laser energy between the droplets inside the cloud. Once such events are captured, interesting studies in the area of aerosol–cloud interactions (i.e., the indirect effects of aerosol radiative forcing), such as studies of the influence of aerosols in the subcloud layer (including those in the boundary layer) on the time evolution of cloud structures aloft, are planned in future work. In an attempt to investigate the time evolution of the nighttime boundary layer and residual layer (a layer formed in the postsunset time because of the settling of aerosol particles, which are lifted into the atmosphere due to convective activity during daytime) over the ex-

1294


VOLUME 25

FIG. 5. Same as Fig. 4, but for range-corrected backscatter intensity, indicating structure and stratifications of nighttime boundary layer. Aerosol plumes close to the surface and the structure of the residual layer can be noted.

perimental station, the lidar was operated throughout the night on 30–31 December 2005. A series of lidar backscatter intensity profiles, commencing around 2130 local time (LT) of 30 December 2005 and continuing to the morning of the next day, at about 0700 LT, were acquired at high spatial resolution at one-minute inter-

vals, and the analysis of this voluminous data was confined to the lowest few kilometers. The raw backscatter intensity profiles thus obtained (up to 300 m with parallel and perpendicular polarization channels of the lidar) are depicted in Fig. 4. The range-corrected backscatter intensity profiles, after treatment for Rayleigh

FIG. 6. Time evolution of the linear depolarization ratio computed from both parallel and perpendicular polarization channels of the lidar on 30/31 Dec 2005.

AUGUST 2008

DEVARA ET AL.

scattering, are mapped in Fig. 5. The time evolution of both the nighttime boundary layer and the residual layer can be seen clearly from the figures. Both the boundary layer and residual layer heights appear larger initially at higher altitudes, decay with the progression of time, and recover and grow quickly from sunrise onward. Another interesting feature that can be witnessed from the figure is the presence of aerosol plumes at certain epochs after midnight. These aerosol plumes are considered to be caused primarily by near-ground temperature inversions. Such temperature inversions, which lead to the formation of stable layers and the subsequent trapping of aerosols during winter months, have been reported in the literature (Devara and Raj 1991). More detailed study of such events in association with atmospheric stability analysis is also planned in future work. Figure 6 displays the time evolution of the linear depolarization ratio (LDR) observed during the night of 30/31 December 2005. Smaller LDR values in the surface layer almost from midnight to early morning hours indicate relatively more isotropic aerosol particles than in the nighttime boundary layer and aloft over the experimental site. The high LDR values throughout the height region from the start of daytime can also be seen clearly. As further observations from this unique facility are made available, additional interesting research, particularly in the area of aerosol– cloud–climate interactions, is planned in future work. Acknowledgments. We thank the Governing Council, Director, and Lidar & Radiometric Group of IITM, Pune, for their support and help. Thanks are due to TejPaul and his team of M/s Foretech Systems Pte. Ltd., Singapore for their assistance in the design and fabrication of the system as per our scientific research goals. The encouragement from the organizers of the ISTP7 and the valuable constructive comments of the anonymous reviewers on the original manuscript of the paper are also gratefully acknowledged. REFERENCES Beninston, M., J. P. Wolf, M. Beniston-Rebetez, H. J. Kölsch, P. Rairoux, and L. Wöste, 1990: Use of lidar measurements and numerical models in air pollution research. J. Geophys. Res., 95, 9879–9894. Bodhaine, B. A., N. B. Wood, E. G. Dutton, and J. R. Slusser, 1999: On Rayleigh optical depth calculations. J. Atmos. Oceanic Technol., 16, 1854–1861. Charlson, R. J., and J. Heintzenberg, Eds., 1995: Aerosol Forcing of Climate. John Wiley, 416 pp. Devara, P. C. S., and P. E. Raj, 1991: Study of atmospheric aerosols in a terrain-induced nocturnal boundary layer using bistatic lidar. Atmos. Environ., 25A, 655–660.

1295

——, R. S. Maheskumar, P. E. Raj, G. Pandithurai, and K. K. Dani, 2002: Recent trends in aerosol climatology and air pollution as inferred from multi-year lidar observations over a tropical urban station. Int. J. Climatol., 22, 435–449. ——, P. E. Raj, G. Pandithurai, K. K. Dani, R. S. Maheskumar, 2003: Relationship between lidar-based observations of aerosol content and monsoon precipitation over a tropical station, Pune, India. Meteor. Appl., 10, 253–262. Fernald, F. G., 1984: Analysis of atmospheric lidar observations: Some comments. Appl. Opt., 23, 652–653. Hansen, J., M. Sato, R. Ruedy, A. Lacis, and V. Oinas, 2000: Global warming in the twenty-first century: An alternative scenario. Proc. Natl. Acad. Sci. USA, 97, 9875–9880. Holm, E. V., 2004: Lidar data applications in numerical weather prediction. Proc. 22nd Int. Laser Radar Conf. (ILRC), Matera, Italy, European Space Agency, 631–634. Houghton, J. T., Y. Ding, D. J. Griggs, M. Noguer, P. J. van der Linden, X. Dai, K. Maskell, and C. A. Johnson, Eds., 2001: Climate Change 2001: The Scientific Basis. Cambridge University Press, 881 pp. Kamineni, R., T. N. Krishnamurti, R. A. Ferrare, S. Ismail, and E. V. Browell, 2003: Impact of high-resolution water vapor cross-sectional data on hurricane forecasting. Geophys. Res. Lett., 30, 1234, doi:10.1029/2002GL016741. Klett, J. D., 1981: Stable analytical inversion solution for processing lidar returns. Appl. Opt., 20, 211–220. Lenschow, D. H., B. B. Stankov, and L. Mahrt, 1979: The rapid morning boundary layer transition. J. Atmos. Sci., 36, 2108– 2124. Mahrt, L., 1985: Vertical structure and turbulence in the very stable boundary layer. J. Atmos. Sci., 42, 2333–2349. McCormick, M. P., and Coauthors, 1993: Scientific investigations planned for the Lidar In-space Technology Experiment (LITE). Bull. Amer. Meteor. Soc., 74, 205–214. Mishchenko, M. I., L. D. Travis, R. A. Kahn, and R. A. West, 1997: Modeling phase functions for dust-like tropospheric aerosols using a shape mixture of randomly oriented polydisperse spheroids. J. Geophys. Res., 102, 16 831–16 847. Ramanathan, V., R. D. Cess, E. F. Harrison, P. Minnis, B. R. Barkstrom, E. Ahmad, and D. Hartmann, 1989: Cloudradiative forcing and climate change: Results from the Earth Radiation Budget Experiment. Science, 243, 57–63. Sasi, M. N., 1994: A reference atmosphere for the Indian equatorial zone. Indian J. Radio Space Phys., 23, 299–312. ——, and K. Sen Gupta, 1986: A reference atmosphere for the Indian zone from surface to 80 km. Vikram Sarabhai Space Centre Space Physics Laboratory Scientific Rep. SPL: SR006:85, 85 pp. Sassen, K., 1991a: Aircraft-produced ice particles in a highly supercooled altocumulus cloud. J. Appl. Meteor., 30, 765–775. ——, 1991b: The polarization lidar technique for cloud research: A review and current assessment. Bull. Amer. Meteor. Soc., 72, 1848–1866. Schotland, R. M., K. Sassen, and R. J. Stone, 1971: Observations by lidar of linear depolarization ratios by hydrometeors. J. Appl. Meteor., 10, 1011–1017. Sivakumar, V., Y. Bhavanikumar, P. B. Rao, K. Mizutani, T. Aoki, M. Yasui, and T. Itabe, 2003: Lidar observed characteristics of the tropical cirrus clouds. Radio Sci., 38, 1094, doi:10.1029/2002RS002719.