next up gif contents
Next: Instrumentation Up: Analysis of Volume Imaging Previous: Contents
Return to the Publications. Return to the Index.


This study introduces methods for measuring convective boundary layer (CBL) mean depths, cloud base altitudes, cloud top altitudes, cloud coverages, and vertical profiles of the horizontal mean wind using data obtained with the University of Wisconsin Volume Imaging Lidar (VIL) [1]. The study uses the VIL data recorded during the First ISLSCP (International Space Land Surface Climatology Project) Field Experiment (FIFE) [2] on the Konza Prairie Natural Area near Manhattan, Kansas, during 1987 and 1989. A primary goal of the ISLSCP was to study problems in scaling from point measurements to the scale of satellite pixels.

Most atmospheric models deal with a mesoscale spatial resolution, but the meteorological data is traditionally measured with point measuring instruments. While these instruments are becoming more accurate, sampling problems still remain. Convection scale atmospheric inhomogeneities cause low-frequency variations in observations. The averaging time interval cannot always be extended enough to reduce these variations without losing information about changing atmospheric conditions. The point measurements can be biased by local forcings due to soil properties and topography. In addition, the bias error can be corrected only by increasing spatial sampling. Therefore, there is no guarantee that a precise point measurement represents accurately the mesoscale phenomenon. Due to these sampling problems, direct observations over mesoscale areas are required.

A lidar is an excellent tool for a wide variety of atmospheric boundary layer (ABL) studies. Each lidar backscatter profile typically contains 100 to 1000 sample points, which effectively reduces statistical variations in measured variables. Several lidar techniques are available for measuring the lower troposphere. A commonly used technique is to detect Rayleigh and Mie backscatter from the atmospheric molecules and aerosols [3]. This technique reveals aerosol structures and optical properties of the atmospheric boundary layer. Sasano et al.[4] used this method to investigate diurnal variations in mixed layer depths. In another study, Uthe [5] reported measurements of aerosols, clouds, and precipitation elements over land and ocean using mobile backscatter lidars. Platt and Takashima[6] studied liquid water content and droplet mode radius retrievals from a CO lidar signal by applying a model describing a relation between the backscatter-to-extinction ratio and droplet mode radius. A lidar has also been used to gather evidence of gravity waves produced by convection in substantial wind shear conditions [7,8].

Scanning improves the sampling coverage and information content of the lidar signal. Eloranta et al. [9] used a lidar pointing at three different directions in alternating shots to reveal wind profiles. Kolev et al. [10,11] used the same technique to investigate the atmospheric boundary layer over urban areas. Installing the lidar in an aircraft also improves sampling coverage. Melfi et al. [12] and Flemant et al. [13] studied the atmospheric boundary layer structure over land and ocean measuring momentum, heat, and moisture fluxes to determine their effect on the evolving structure of the ABL.

Researchers have applied spectroscopic methods to customize lidars for accurate remote sensing of chosen physical properties of the atmosphere. Doppler-lidar has been used to study air mass flows [14,15]. The Doppler-lidar detects a shift in the backscatter signal wavelength caused by wind. The method has been successfully used to examine mountain valley flows, low-level flows leading to thunderstorm formation, dry-line front structure, dynamics of smoke columns from a forest fire, and structure of the nocturnal drainage winds in valleys. The Doppler lidar is important for investigating the turbulence in the atmospheric boundary layer[16].

Uthe [17] used two lidars with different wavelengths to study the feasibility of remote evaluation of dust cloud types. The cloud types were determined by examining the scattering and extinction properties of clouds at different wavelengths. Raman scattering provides a more practical method to separate different scattering constituents of the atmosphere, since Raman scattering occurs at individual wavelength shifts for each atmospheric scatterer. The NASA Goddard Space Flight Center's Raman lidar is capable of measuring water vapor content simultaneously with Rayleigh and Mie scattering[18]. The combined Raman/Rayleigh-Mie lidar along with the High Spectral Resolution Lidar [19] can separate the Rayleigh and Mie scattering without using an assumed relation between the backscatter and extinction cross sections[20]. Since the Raman backscatter cross section is small compared to Mie and Rayleigh scattering, it is important to suppress the solar background radiation. A daytime Raman lidar is usually based on the use of wavelengths between 250 nm and 300 nm, where atmospheric ozone filters out the solar radiation almost completely [21,22,23]. This technique requires correction for differential absorption caused by ozone. A Raman lidar can be also used to measure atmospheric temperature [24]. A more sophisticated lidar system, however, can provide better temperature measurements by detecting pure rotational Raman backscattering of nitrogen and oxygen molecules [25,26].

Mobile lidar systems have proven to be useful for remotely measuring air pollution and its transport. For example, Mitsev et al.[27] reported investigations of mass-concentration fields near a coal power plant. Fredriksson et al. [28] introduced a mobile system consisting of a Mie lidar, differential absorption lidar (DIAL), and Raman lidar for air pollution studies. The DIAL measures the atmosphere using two wavelengths: one on the absorption peak and another off the absorption peak of a selected atmospheric constituent. The wavelength-dependency of backscatter is smooth, while the absorption is highly wavelength-selective. Thus, the absorption difference between backscattered signals on and off the absorption peak reveals the quantity of the absorbing media. Aldén et al. [29] and Edner et al. [30,31] made DIAL measurements of atmospheric mercury, SO, and NO. They reported Hg-concentration levels up to 3gm near mercury distilling plants and banks of roasted cinnabar (HgS) ore. Browell et al. [32] used an aircraft-based DIAL to map ozone levels in the lower troposphere.

Fluorescence lidar provides another powerful tool for air pollution studies. It is based on the detection of organic, fluorescent material from the atmosphere. When stimulated by a short-wavelength laser pulse, fluorescent materials emit light with individual spectrum that enables identification of different fluorescent specimens. The technique can be used for long-range identification of biological particulates and cyclic aromatic hydrocarbon derivatives [33]. This has potential applications for monitoring environmental pollution, industrial waste emission, agricultural insect control, and illicit chemical processes. Rowland and Konrad [34] used a fluorescence lidar to study the dispersion and transport of smoke from an industrial source. Schuster and Kyle [35] proposed doping smoke with artificial fluorescent colors to create a controllable calibration target inside the plume. Using a color which does not exist in nature, the measurement range can be extended far from the source where diffusion and turbulence have already diluted the plume. Uthe et al. [36] successfully demonstrated this method for tracing diffuse plume sources.

Most of the previously mentioned lidars are relatively large instruments mounted on vans, trucks, semi-trailers or freight aircrafts. Recent advances in laser, detector, and computer technologies have enabled development of compact, portable lidars for applications where size requirements are critical. Spinhirne and Scott [37] introduced a table-top lidar system capable of measuring clouds and aerosols up to 20 km. This system consists of a diode-pumped Nd:YLF laser, 20 cm Cassegrain telescope, solid state photon counting module, portable PC computer with a multichannel scaler card, and power sources for the laser and photon counting module. One person can easily handle the system and it fits on a table top.

This study uses data obtained with the University of Wisconsin Volume Imaging Lidar. The unique scanning capability of the VIL provides great advantages over other lidars in atmospheric boundary layer studies. The VIL measures three-dimensional aerosol backscatter in large (>100 km) volumes with good spatial and time resolution. Currently, no other lidar can compete with the VIL in the combination of coverage, spatial resolution, and scanning speed. Section 2 discusses the characteristics of the VIL in detail.

The VIL provides information for meteorological investigations of the atmospheric boundary layer. Elevation angle scans of the VIL allow inspection of convective boundary layer mean depths, cloud base altitudes, and cloud top altitudes [38,39,40,41].

The three-dimensional aerosol backscatter fields provide information about the size distribution and dynamic properties of large eddies, which can be used in developing large eddy simulation models [42]. The fast scanning speed of the VIL ensures that most of the aerosol structures which appear in one volume scan are coherent in the next scan. This enables mean wind [43,44,45] and divergence[46] calculations by tracking aerosol structures.

This thesis presents a comprehensive study of the measurements of the convective boundary layer mean depth, cloud base altitude, cloud top altitude, cloud coverage, and vertical profiles of the horizontal mean wind from the VIL data. The methods for retrieving structural parameters of the atmospheric boundary layer and estimating geometrical properties of clouds are presented in section 3. The wind profiling method is introduced in section 4. Refinements are made to a correlation technique which calculates vertical profiles of horizontally averaged mean winds. The efficiency of the technique is optimized and a comprehensive error analysis is accomplished. To ensure the reliability of the VIL results, they are compared with measurements from weather stations, radiosondes, aircraft-based instruments, and satellites.

The studied methods provide unique area-averaged measurements with scales similar to satellite pixel size and a time resolution close to point measurements. The estimates determined from the data typically represent averages over a 70 km scanning area. When using a traditional point measuring instrument, similar coverage would require a line average over thousands of kilometers [47]. In addition, a 15-meter range resolution along with a 0.33--0.5 angular step size provide local observations with better than a 15 m 130 m horizontal area resolution. The VIL data offers a scalable link between point and mesoscale observations.

The automatic methods developed in this thesis can be modified to use with other scanning lidars, although the performance of these methods varies with scanning properties, repetition rate, and signal quality. The area-averaged results open new possibilities for more comprehensive studies of the boundary layer behavior. The VIL results from the FIFE program provide unique, area-averaged measurements to support other studies of the convective boundary layer [42,48]. Due to the large number of VIL results from the FIFE program, only brief summaries of the results are presented in sections 3.3 and 4.3.4. Complete sets of the results are available via a FIFE user database [49,50].

This thesis summarizes a three-year study on the analysis of the VIL signals. Some results of the study were presented in the Sixteenth International Laser Radar Conference [43], Cambridge, Massachusetts, 1992, and Optical Remote Sensing of the Atmosphere Topical Meeting[44], Salt Lake City, Utah, 1993. This study has also produced manuscripts for two journal articles [51,52], which were revised at the time this thesis was submitted for printing.

next up gif contents
Next: Instrumentation Up: Analysis of Volume Imaging Previous: Contents
Return to the Publications. Return to the Index.

Antti Piironen
Tue Mar 26 20:53:05 CST 1996