TITLE: Airborne Lidar-Derived Boundary Layer Depth Data AUTHORS: Kenneth Craig Kenneth J. Davis (PI) Graduate Research Assistant Associate Professor Department of Meteorology Department of Meteorology The Pennsylvania State University The Pennsylvania State University 415 Walker Building 512 Walker Building University Park, PA 16802 University Park, PA 16802 DATA SET OVERVIEW: This dataset contains PBL depths retrieved from airborne lidar backscatter data collected during the IHOP experiment, 13 May to 25 June 2002. The data are in ASCII format. Data are available from IHOP missions flown on the the following days: May 19,20,21,25,27,28,29,30,31 June 03,06,07,09,14,16,25. INSTRUMENT DESCRIPTION: Lidar data used to produce this dataset were provided by the principle investigators of the airborne lidar systems that participated in IHOP. For more information about the lidar instruments, data, or platforms, consult the appropriate lidar metadata information, or contact the principle investigators below. DLR LIDAR (flown on DLR Falcon) LEANDRE LIDAR (flown on P-3) PI: Gerhard Ehret and Christoph Kiemle PI: Cyrille Flamant DLR Oberpfaffenhofen CNRS Service Aeronomie Institut fuer Physik der Atmosphaere/Lidar Universite Pierre et Marie Curie Muenchnerstr. 20 Tour 15 Couloir 15-14 82234 Wessling, Germany 4 Place Jussieu gerhard.ehret@dlr.de 75252 Paris Cedex 05, France christoph.kiemle@dlr.de cyrille.flamant@aero.jussieu.fr LASE LIDAR (flown on NASA/DC-8) PI: Ed Browell MS 401A NASA Langely Research Center Hampton, VA 23681, USA e.v.browell@larc.nasa.gov A PBL depth is determined for each vertical backscattering profile obtained from the lidar. Each profile represents 5 or 6 second averages; therefore, the horizontal (i.e. along flight track) resolution of this dataset is 700-1000 m, depending on platform and airspeed. The vertical resolution of this dataset is equal to the native lidar vertical resolution (15 or 30 m). No interpolations are made between the discrete lidar range gates. This dataset is divided into flight transects consisting of 50 to 500 profiles each. A transect typically consists of a single linear flight segment, often oriented north-south or east-west. In many cases, aircraft turns were eliminated from the lidar dataset before PBL depth retrieval. When aircraft turns (defined as aircraft roll greater than 5 degrees) occur in the lidar data, the retrieved PBL depth is objectively screened. DATA PROCESSING: 1. Method PBL depths were retrieved from downward-pointing airborne lidar backscatter data using the wavelet method of Davis et al. (2000). Aerosols lofted from the surface by convective eddies are trapped in the PBL by a potential temperature inversion at the PBL top. A sharp contrast typically exists between the aerosol-laden PBL and the cleaner free atmosphere. The wavelet technique objectively determines the location of this boundary. As described in Davis et al. (2000), the wavelet dilation (size of the wavelet) can be varied to retrieve atmospheric structures at various scales. When possible, the wavelet dilation is objectively determined for each profile as the dilation that produces the maximum wavelet variance, since this defines the scale of the dominant structure in the backscatter profile. All cloud-free backscatter profiles contain a ground spike, an abrupt increase in backscatter due to the ground. The location of the ground spike is determined from a gradient approach, as the ground spike usually represents the strongest gate-to-gate backscatter gradient in a cloud-free profile. The ground relative PBL depth is calculated as the difference between the height of the retrieved PBL top and the elevation of the ground spike. 2. Quality Control This dataset was objectively and subjectively screened for cloud contamination. Clouds strongly attenuate the lidar signal and produce strong backscatter, which causes the wavelet method to produce erroneous PBL depths (i.e., cloud top is retrieved). Because of the attenuation, the ground spike is not present and the cloud represents the strongest backscattering gradient in the profile. Therefore, PBL depths from cloudy profiles can be screened objectively. Data are subjectively screened when clouds that slip past the objective screening process affect the PBL depth retrieval. When multiple boundaries are present, or when the backscatter contrast between the PBL and the free atmosphere is weak, it sometimes becomes necessary to limit the range of wavelet dilations allowed, or to set the dilation to a fixed value for a particular transect to prevent erroneous retrievals. The decision to fix or limit the dilation is made for each flight transect, based on a visual inspection of how the retrieved PBL depth fits the backscatter data for that transect. Each transect of PBL depth data was subjected to visual comparison with the corresponding lidar backscatter data to ensure that the PBL depth retrieval followed the PBL top in the backscatter data. In many instances, the wavelet method is robust, and can faithfully retrieve PBL depths from copious amounts of data with little or no subjective screening. However, in some instances, the presence of multiple layers near the PBL top produces erroneous retrievals, regardless of the selected wavelet dilation. In other instances, the contrast between PBL and free atmosphere backscatter is simply too low to objectively (or even subjectively) detect the PBL top. In cases where the wavelet method obviously produced an erroreous PBL top, those erroneous data points were manually screened. In no instance was a retrieved PBL depth arbitrarily changed to a value that "appeared" to coincide with PBL depth seen in the backscatter data. If the wavelet method produced a PBL depth that was obviously in error, that value was removed from the dataset, even if the PBL top was visually apparent in the backscatter data. DATA FORMAT Data files are presented in ASCII format. Each file represents one flight segment. Files are named PLATFORMmmddlegiizi.txt, where PLATFORM = the lidar platform from which data were derived (LEANDRE, DLR, or LASE), mm = month, dd = day, leg = leg number. The zi indicates that the dataset contains PBL depth. All data are for the year 2002. For example, DLR0525leg05zi.txt contains PBL depths derived from transect 5 of DLR data obtained May 25 2002. On the Penn State ABL-group IHOP site, data files are organized by lidar platform, then by date. Each platform directory contains a tar file with all the data for that platform. Each date directory contains individual files for each transect, and a tar file that contains all the files for that date. The first 14 lines of each file are the header. The first line is the number of data points in the file (integer). The next 13 lines are character strings that contain information about the data in the file. The data are presented in tabular format, and contain time, location, PBL depth with respect to mean sea level (zi_msl), and PBL depth with respect to the ground (zi_agl). Bad data points are denoted by -999.00. Sample programs in FORTRAN and IDL are available on the ftp site (read_public_zi.f90, read_public_zi.pro) as templates to extract data from these files. Below is part of a sample file to illustrate the file format. 313 BOUNDARY LAYER DEPTH RETRIEVAL IHOP 2002 EXPERIMENT File: LEANDRE0519leg07zi.asc Created: Fri Jan 16 08:14:12 EST 2004 Platform: LEANDRE Lidar Date: 05/19/2002 Time: 18:59:49 - 19:26:01 Number of Data Points: 313 Average Aircraft Speed (m/s): 122.358 Starting Aircraft Altitude (m MSL): 4332.16 Ending Aircraft Altitude (m MSL): 4328.83 PI: Dr. Ken Davis (davis@met.psu.edu). 503 Walker Building, University Park, PA 16802-5013 TIME LATITUDE LONGITUDE ZI_MSL ZI_AGL 18:59:49 37.380 -99.823 1770.00 1305.00 18:59:54 37.380 -99.830 1635.00 1170.00 18:59:59 37.380 -99.837 1605.00 1125.00 19:00:04 37.380 -99.844 1755.00 1260.00 19:00:09 37.380 -99.850 1560.00 1065.00 REMARKS None. REFERENCES Davis, K. J., N. Gamage, C. Hagelberg, D. H. Lenschow, C. Kiemle and P. P. Sullivan, 2000: An objective method for determining atmospheric structure from airborne lidar observations. J. Atmos. Oceanic Tech., 17, 1455-1468.