ARM/GCIP NESOB 1996 Cloud Data Sets 1.0 General Description The ARM Cloud and Radiation Testbed (CART) cloud height dataset is one of several datasets provided for the GEWEX Continental-Scale International Project (GCIP) Near-Surface Observing Period (NESOB) 1996 project. This dataset contains cloud height data collected using the ARMCART SGP Belfort Laser Ceilometer (BLC) Model 7013C and the ARMCART SGP MicroPulse Lidar (MPL) located at station E13: Lamont Central Facility 1 (CF1). This dataset contains data from these two instruments located within the NESOB 1996 domain (94.5 W to 100.5 W longitude and 34 N to 39 N latitude) for the NESOB 1996 time period (01 April 1996 through 30 September 1996). These data were not quality controlled by the University Corporation for Atmospheric Research/ Joint Office for Science Support (UCAR/JOSS). The cloud base heights reported by the Belfort Laser Ceilometer currently have a vertical offset of undetermined origin which results in a positive bias of approximately 110 meters. For more information on this bias, see section 2.2. 2.0 Detailed Data Description 2.0.1 ARMCART SGP Belfort Laser Ceilometer (BLC) Model 7013C Data Description The data in this dataset were formed by extracting 30 minute cloud height data from 60 second BLC data in netCDF format. A zero for the first cloud layer signifies no clouds detected within the limit of the instrument field of view and range. A cloud ht of zero for the second and third cloud layers indicates an unknown cloud height (i.e. the ceiliometer can't see through lower clouds). These zeros are reported as "0.00000 U" in the data. The following are descriptions of the algorithms used by ARMCART to produce the Belfort Laser Ceilometer (BLC) Model 7013C Data. A complete data description, including figures can be found at the ARM web site at http://www.arm.gov/docs/instruments/static/blc.html. The Model 7013C Ceilometer field unit is a self-contained, ground-based, optical, active remote sensing instrument with the ability to detect and process several cloud-related parameters. These parameters include cloud height, extinction coefficient, cloud layers and time/date reference information. The ceilometer system detects clouds by transmitting pulses of infrared light vertically into the atmosphere. The receiver telescope detects scattered light from clouds and precipitation. The BLC measures the following quantities between the heights of 15 and 7350 m directly above Mean Sea Level: Base height of lowest cloud detected Base height of second-lowest cloud detected Base height of third-lowest cloud detected The base height of the second- and third-lowest clouds are not always reported because they can be optically obscured by the lower clouds. In addition, visibility is reported and could be considered a secondary variable measured. The values for visibility are a qualitative indication of vertical visibility. A precise interpretation is difficult, because algorithm details have not been provided by the manufacturer. List of Components The BLC field unit contains all the electronics necessary to collect and process information about cloud formations, height, and other conditions. The collected data is then processed and transmitted to the Site Data System (SDS). The BLC components include the following: Transmitter Telescope Optical Assemblies Heaters and Fans Receiver Printed Circuit Board (PCB) PC104 Input/Output (I/O) PCB Receiver Telescope Direct Current (DC) Power Supply Transmitter PCB PC104 Central Processing Unit (CPU) PCB Digital Signal Processor (DSP) PCB The ceilometer system includes the field unit and primary display unit. Description of System Configuration and Measurement Methods Light pulses generated by a laser diode are in the near-infrared range. The ceilometer uses a fiber-optic cable to transfer the output of the Gallium Arsenide laser diode to the focal point of a Newtonian telescope. The light energy is projected upward by the mirror in the transmitter telescope. The receiver telescope is aligned with the transmitter telescope for detecting light backscattered from any object in the transmitted beam. The beam divergence of both telescopes is less than 3 milliradians. The amount and quality of returned light will depend on the density of clouds, rain, snow, or fog. The size of water droplets in the atmosphere will also affect the return signals. Each data acquisition cycle consists of 5120 pulses of the laser diode. Immediately following the first pulse, the receiver telescope begins detecting the light backscattered from the atmosphere. These pulses are 150 ns wide and occur at 1.024 millisecond intervals. The data acquisition cycle can be repeated at a user-selected rate, but the default rate is every 30 seconds. In the first 5.12 seconds of a 30-second cycle, it sends out 5120 pulses and detects the backscatter. From 5.12 seconds until 10-12 seconds it does number crunching to filter out noise and determine cloud presence and then dumps out its data stream. From 13 seconds until the end of the 30-second cycle, it does nothing. The time stamp within the data stream reflects two seconds before the beginning of the firing of the pulses, hence, the beginning of the cycle. The return signal for each transmitted pulse is analyzed, and information is stored in range cells that represent 7.6-meter (25-foot) increments. When the pulses are completed, the information stored is further analyzed and converted to data for transfer. The two covers of the BLC protect the electronics from environmental conditions. The two telescopes are mounted on the back of the plate and aligned for maximum efficiency. A glass cover over the telescopes is slanted to allow rain and melted snow to run off. Any snow that falls on the telescopes is quickly melted by the internal heating system that sufficiently maintains the temperature inside the unit during cold weather. The transmitter telescope transmits a pulsed laser-beam vertically into the atmosphere. Under the direction of the PC104 CPU,the DSP PCB sends 5120 trigger pulses to the transmitter PCB so that a pulsed laser-beam is generated for each data acquisition cycle. The receiver telescope detects backscattered energy and converts it to digital data. This data is continuously processed by the DSP PCB in realtime during the 5120 pulses. After each pulse fired, the data is stored in range cells according to the time received after the transmitted pulse. Upon completion of the 5120th pulse, the collected data is analyzed, processed, and transmitted to the SDS. Each data acquisition cycle consists of 5120 pulses of the laser diode. These pulses are 150 ns wide and occur at 1.024-millisecond intervals. Immediately following the first pulse, the receiver telescope begins detecting the light backscattered from the atmosphere. The data acquisition cycle can be repeated at a user-selected rate with the default stored in the system configuration of the ceilometer. Selectable rates range from 30 seconds up to 3600 seconds. The Data Acquisition Cycle During normal operation, the ceilometer performs a data acquisition cycle every 30 seconds. Each data acquisition cycle has a duration of 5.24 seconds and creates four data arrays in Random Access Memory (RAM) on the DSP PCB. These four data arrays are the summations of data for Phase 0, Phase 1, Phase 2 and Phase 3. A data acquisition cycle consists of 5120 consecutive phase cycles, each of 1.024 millisecond in duration. During each phase cycle, the laser diode is pulsed for 150 ns in Phase 0. Then, in Phase 1, the input data-comparator is given a positive offset voltage of +6.2 mV. In Phase 2, the input data-comparator is given a negative offset voltage of -6.2 mV. Finally, in Phase 3, the data comparator is given no offset voltage. The purpose of Phase 1 and Phase 2 during each phase cycle is to convert the signal-to-noise (S/N) ratio of the received signal into a voltage. The four phases are described as follows: Phase 0: Received echo from a fired laser. Phase 1: Received background with +6.2 mV offset at the comparator in place of a laser echo signal. Phase 2: Received background with -6.2 mV offset at the comparator in place of a laser echo signal. Phase 3: Received background with no laser echo and no offset. During each of the four phases, 1024 bits of data (range cells) are collected at a 20-MHz rate and then added to the running sum (accumulated) for that particular phase. At the end of the entire data acquisition cycle, four 1024 by 16 arrays of accumulated data are stored in RAM on the DSP PCB. The 20 MHz sampling rate (50 ns) provides a height resolution of 7.6 meters (25 feet). By sampling data at 50-ns intervals, the laser light has time to travel a total of 15 meters (50 feet). These 15 meters consists of 7.6 meters up into the atmosphere and 7.6 meters to return to the BLC. Optical System The transmitter optics and receiver optics are nearly identical. Each telescope consists of a concave mirror mounted at the bottom of the telescope tube. In the transmitter telescope, a fiber-optic cable conducts laser light from the laser diode down towards the transmitter mirror. In the receiver telescope, a fiber-optic cable conducts laser light from the receiver mirror up to the Avalanche Photo-Diode (APD) on the receiver PCB. Each telescope consists of a 21.6-cm diameter tube case with a 20.3-cm diameter concave mirror mounted in the bottom. At the top of the tube, a "spider" support-bracket is mounted to hold one end of each fiber-optic cable near the focal point of its associated mirror. The ends of the two fiber-optic cables are aligned for precise parallel collimation of the two telescopes. The other end of each fiber-optic cable is optically coupled to the transmitter optics assembly and receiver optics assembly. In the case of the transmitter laser diode, the light is focused into the fiber-optic cable for optimum transfer of energy into the telescope. The temperature of the transmitter laser diode is monitored and a Peltier cooler is used to maintain a proper temperature range. If the laser diode temperature exceeds predetermined limits, voltage to the laser diode is temporarily reduced to 70 V, but is restored when the laser diode returns to an acceptable temperature. Another safety feature is the monitoring of the frequency and duration of the trigger pulses. If acceptable limits are exceeded during a data acquisition cycle, voltage to the laser diode is reduced to 0 V for the next data acquisition cycle. This feature protects the laser diode from damage and to ensure compliance with eye-safety regulations. In the receiver telescope system, the light from the fiber-optic cable is focused into the avalanche photo-detector (APD). The receiver lens system has two lens with an infrared light filter located in between. The first lens is used to collimate the light into the filter. The filter rejects light wavelengths outside the desired range and passes the desired light into the second lens. The second lens focuses the received light into the APD. Using the filter increases the signal to noise ratio of the system by rejecting unwanted light wavelengths. The effects of sunlight and other sources of possible optical noises are reduced. Firmware Overview The entire operation of the BLC is controlled by a master microprocessor located on the PC104 CPU PCB. This microprocessor, also referred to as AM286LX-16 or PC104 CPU, executes the firmware programmed into the 512KB Electrically Programmable Read Only Memory (EPROM) in socket S0. The PC104 CPU also uses the DR DOS 5.0 Operating System. The transmitter PCB has a microcontroller that acts as a slave to the PC104 CPU. This microcontroller, also referred to as 80C552 or transmitter CPU, executes the firmware programmed into the 64KB EPROM in U28. The PC104 CPU performs numerous software tasks and hardware control activities that enable the BLC to detect and process cloud-height information. One of these tasks is communicating with and controlling the transmitter CPU that, in turn, directly controls the hardware on the transmitter PCB. Laser firing, data-acquisition, digital-signal-processing, system diagnostics and interfacing to an external user or system are just some of the other tasks performed by the PC104 CPU. The transmitter CPU acts as a slave to the PC104 CPU. Therefore, one of the primary tasks of the transmitter CPU is to communicate (via serial port COM3) with the PC104 CPU. Some of the other tasks performed by the transmitter CPU include controlling the supply voltages for the laser diode and APD, controlling the Peltier cooler and performing transmitter PCB diagnostics. Processing Received Signals Stored data in the four arrays must be processed mathematically to determine what signal power actually entered the receiver telescope at each range cell (7.6 meter increment). The received power data is required for input to a Klett lidar inversion program. The data for each phase array contains information to be used in processing the return signals. Signal plus noise data, noise data with offsets, and background data make it possible to extract the desired signal from the raw data. Coupled with gain control for altitude and established factors from calibration, the resulting information will indicate how much light was backscattered by aerosols over a range of 15 meters to 7350 meters of altitude. The phase zero array consists of returned signal plus noise plus some breakthrough. The multiplier breakthrough (contamination from the swept gain waveform) is identical in Phase 0 and Phase 3. Subtracting Phase 3 from Phase 0 would therefore remove the multiplier breakthrough. However, a direct "range cell for range cell" subtraction would cause noise level to double, reducing detectability. Since multiplier breakthrough is known to be a low frequency effect, a filter is able to extract a good, low noise estimation of the true multiplier breakthrough. It is this curve that is subtracted from Phase 0 to remove the effect of multiplier breakthrough. An additional breakthrough signal occurs due to some electrical coupling of laser firing energy into the Receiver preamplifier. This breakthrough appears as ringing in the first few range cells of Phase 0. A ring calibration procedure takes advantage of the fact that the ringing signal is stable and virtually independent of detector gain. The ring signal data is recorded and stored in EPROM during calibration. During normal operation, the ring signal is removed by subtracting the stored data from present data. The array that results after subtracting the two breakthrough effects is called the signal count array. The difference between Phase 1 and Phase 2 data is used to determine a conversion factor from comparator crossing counts to actual signal volts input to the comparator. The conversion factor array contains the ratio of volts to counts for each range cell. The signal volts array is, therefore, produced by multiplying the signal counts array by the conversion factor array. The volt array is then followed by an adaptive window filter. The adaptive window filter applies an averaging window size based on a signal-to-noise ratio (S/N) of range cells. Range cells with a sufficiently high S/N ratio get an averaging window size of one (no averaging). As S/N ratio decreases, averaging window size for associated range cells increases. Noise level for S/N ratio determination is obtained as a function of average conversion factor at high range and a swept gain attenuation function. The computed noise level for each range cell is stored in a noise level array for use by the adaptive window filter and the squelch. A squelch is applied to reject range cells having a S/N ratio, after adaptive window averaging, that is too low for use by the Klett program. From the receiver front end to the comparator, the received signalis multiplied by a swept gain function to increase the dynamic range of the receiver. The effect of the swept gain function must be inverted so that the level of signal input to the receiver may be determined. To accomplish this, the signal volt array is multiplied by the reciprocal of the true, compound, swept gain function array. The pre swept gain volts array is the result. After swept gain inversion, two factors are apllied to obtain the actual signal power that struck the receiver telescope mirror. One factor is the system gain factor, which is established during ceilometer calibration using a system gain evaluator. The other factor is in an optical crossover array, X(R), which gives the fraction of beam overlap between the transmitter telescope and the receiver telescope as a function of range. An array, P(R), is the result of application of these two factors. P(R) is the actual power of echo light hitting the Receiver Telescope as a function of range. Finally, a simple function converts P(R) to an array S(R) that is actually used as input to the Klett lidar inversion program. Assessment of System Uncertainties for Primary Quantities Measured The ceilometer can detect up to three cloud layers at a time depending upon whether its beam is able to penetrate the previous layers and whether the backscattering is able to penetrate back through the lower layers and reach the detector. Because of this, the first layer indicated is usually the most accurate of the representations. Detection of multiple cloud layers may also be indicative of a broken or scattered lower deck, with detection of an upper cloud layer through gaps or holes in the lower cloud deck. However, because the ceilometer only actively collects backscattered photons for about 5 seconds of every 30 second measurement period, isolated clouds may be missed by the instrument. The intensity of backscattered light will be dependent on the density of clouds, rain, snow, or fog. The size of water droplets in the atmosphere will also affect the return signals. Additionally, the intensity of the return signal follows the inverse-square law. Therefore, in harsh sky conditions, the higher the clouds are in the sky, the higher the chance the clouds could be missed by the Ceilometer. Users of the data should be aware of these factors and that the ceilometer has at times indicated clear skies when in fact there were clouds present. Site operators said that when the cloud cover became more dense, the ceilometer resumed indicating cloud presence. The sky conditions that increase the probability that the reported cloud information could be inaccurate are described below: Very thin clouds in the presence of sunlight. Scattered or broken cloud layers. Harsh sky conditions such as heavy fog, rain or other precipitation, which blind the detector. A comparison was performed between a Vaisala 25K ceilometer over a 20-day period from 15 September to 5 October 1997. There were enough clouds within a 9-day period to be able to compare cloud heights. The comparison was complicated by the fact that there was about a 5-hour time offset in the VCEIL data and by the fact that the BLC data were contaminated by cloud-height values above 100,000 m. After windowing the data to remove the time offset and ignoring the spurious BLC cloud heights, the two measurements compared well, with both measurements showing the same general shape and the BLC showing clouds heights 100- to 120-m higher than the VCEIL. An example of the comparison can be seen on the BLC web site at http://www.arm.gov/docs/instruments/static/blc.html Description of Observational Specifications BLC specifications are as follows: Dimensions: 152 H X 72.4 W X 65.4 D centimeters Weight: Field Unit 56.5 kg (125 lbs) Shipping 105 kg (230 lbs) Supply voltage: 115VAC/60Hz (U.S. Model) Power Consumption: 930 W Total (800 W Heaters, 130 W Electronics) Height Range: 15 to 7350 meters Height Resolution: 7.6 meters Optical Source: Gallium Arsenide Laser-Diode Wavelength: 910 nm (near-infrared) Peak Optical Output: 15 W (nominal) Optical Pulse Length: 150 ns (nominal) Optical Pulse Repetition Frequency: 976.6 Hz Receiver Optical Detector: Avalanche Photo-Detector (APD) Extinction coefficient variation: +/- 3 dB Update Period: 30 seconds to 3600 seconds in 1-second increments Environmental: Temperature Range (storage): -50 to 65 degrees Celsius Temperature Range (Operational): -30 to 40 degrees Celsius Relative Humidity: 0-100% Wind Speed: Max operational 50 m/s Max survival on concrete base 75 m/s Theory of Operations The BLC uses light wavelengths in the near-infrared range to detect clouds and other weather conditions. The BLC transmits a series of light pulses into the atmosphere and detects, processes and analyzes the returned signals backscattered by the atmosphere. This principle allows the distance to light scattering aerosols, such as a cloud, to be determined. The light detection and ranging (LIDAR) system is similar to Radar in that a pulse of energy is emitted and returned energy analyzed for elapsed time between emitted pulse and received pulses. The major difference is the frequency of emitted energy. Radar uses high range radio frequencies which will reflect from any object that conducts electricity. The LIDAR system uses light energy in the near-infrared region. An additional factor is the effect of light scattering according to the type of reflecting object. Elapsed time between transmission of a light pulse and backscattered energy is used to establish the distance to the reflecting object. The collected data is accumulated and stored in increments representing 7.6 meters and sent to the Site Data System (SDS). Cloud layers are indicated according to their height above the unit. Calibration Theory Because the ARM program relies entirely on original factory calibration procedures for the BLC, which include proprietary information, a description of calibration theory is not available. Evaluation of the performance of the BLC is carried out primarily by comparison of cloud base height data with measurements made by other systems and by an on-line diagnostic procedure identified as "STAT". Calibration History The BLC is factory calibrated by the manufacturer whom, when any of the instrument boards are serviced, also does a recalibration. The command, "STAT" will check the status of the instrument to determine if everything is in normal working mode (see "Belfort Model 7013C Laser Ceilometer: Operating and Maintenance Procedures" by T.J. Peters, May 1995, Pacific Northwest National Laboratory). The BLC has never been calibrated other than when originally purchased from the manufacturer. Contacts The Instrument Mentor at ARMCART is Connor Flynn, PhD. He is a Post-Doctoral Fellow of Associated Western Universities. He can be reached at Pacific Northwest National Laboratory, P.O. Box 999, Richland, WA 99352. Phone: (509) 375-2041, Fax (509) 375-4545, Email: Connor.Flynn@arm.gov The Vendor/Instrument Developer is Belfort Instruments, Baltimore, Maryland. Contact Dave Meenahan or Barry Davis at 410-342-2626. 2.0.2 ARMCART SGP MicroPulse Lidar Data Description The data in this dataset were formed by extracting 30 minute cloud height data from 60 second MPL data in netCDF format. A zero for the first cloud layer signifies no clouds detected within the limit of the instrument field of view and range. A cloud ht of zero for the second and third cloud layers indicates an unknown cloud height (i.e. the ceiliometer can't see through lower clouds). These zeros are reported as "0.00000 U" in the data. The following are descriptions of the algorithms used by ARMCART to produce the MicroPulse Lidar Data. A complete data description, including figures can be found at the ARM web site at http://www.arm.gov/docs/instruments/static/mpl.html. The Micropulse Lidar (MPL) is a ground-based optical remote sensing system designed primarily to determine the altitude of clouds overhead. The physical principle is the same as for radar. A pulse of energy is transmitted and the energy reflected back is measured. From the time delay between the transmitted pulse and the backscattered signal, the distance to the scatterer is infered. Besides real-time detection of clouds, post-processing can also characterize the extent of the tropospheric mixing layer (the planetary boundary layer), or other particle-laden regions. This eye-safe system is designed for continuous operation. Primary Quantities Measured with System The Micropulse Lidar (MPL) has one measurement channel that records backscatter signals in 300 meter range bins, with the lowest valid range bin beginning at 120 meters above ground level, up to 20+ kilometers. The primary quantity obtained from this signal is the real-time reporting of the lowest detected cloud base in meters, obtained from 60 second averages. From the relative backscatter profile, other data products are possible. These include cloud boundaries, multiple cloud decks, and layer boundaries. Detailed Description List of Components The MPL consists of four interconnected subsystems: Data acquisition, control electronics, transmission optics, and detection optics. The MPL consists of four main components: a rack-mounted computer for data aquisition, a dedicated lidar control and power supply, a diode-laser supply of pump radiation, and a shared optics assembly for transmitting the laser pulses and detecting the collected photons. The rack-mounted 486 PC is a Broadax VIC-9000A. It houses a multichannel scalar, model MCS-II-PC-N from Santa Fe Energy Research Corp., which records counts from the detection optics, storing them in 2 microsecond bins. The lidar control and power supply, custom produced by Science and Engineering Services Inc. (SESI), provides controlled voltages to the photon detector and laser energy monitor. It also contains an integrated A/D converter for reporting of vital system parameters to the instrument PC. A Spectra Physics 7300-L3 Diode Module provides 1.00 W of infrared pump radiation from a GaAlAs diode-laser array. The pump radiation is delivered to the transmission optics by fiber optic (Spectra Physics part number 0129-0400). The transmission optics includes the Spectra Physics solid state extended-cavity YLF laser, Model 7960-L3-E, and frequency doubler, Model 7965. These are pumped with the diode laser to produce 523 nm radiation in 10-15 microjoule pulses with 10 ns pulse width at a rate of 2500 Hz. The pulses pass through a depolarizing wedge (CVI ) and expanded to fill the aperture of 8" Celestron telescope from Company 7 Astro-Optics. The detection optics begins with the same 8" Celestron telescope. Very narrow interference filters (.13 nm fwhm) reject most ambient light. Photons are focussed onto the Avalanche photo-diode SPCM-AQ-121 from EG&G Optoelectronics operated in geiger-mode photon counting. Description of System Configuration and Measurement Methods The MPL employs a conventional time-gated incoherent detection scheme. It is a single-telescope design, using an 8-inch catadioptic telescope as both transmitter and receiver. The MPL achieves eyesafe operation by expanding the transmitted beam to 8 inches diameter and by using pulses of microjoule energy. To compensate for the low pulse energy, the pulse repetition frequency (PRF) is high (2500 Hz). Nevertheless, to suppress signal from ambient light, the detector FOV is narrow (100 µrad = 0.006 °). Assessment of System Uncertainties for Primary Quantities Measured The data aquisition software currently collects data in 200 sequential 2-microsecond bins. This aquisition corresponds to a range resolution of 300 meters. An expected modification of the data aquisition software will permit collection of 4 sets of 200 bins, with each set being offset from the previous set by 75 meters. One backscatter profile is recorded for each laser pulse, that is, 2500 times per second. Profiles are summed over 60 seconds, which yields the temporal resolution of 1 minute. Clouds are only searched for in the first 60 bins = 18 km. Presently, only the first detected cloud base is reported, although analysis of the backscatter profile with other algorithms can provide several base heights. A strong point of the MPL is its long-range detection. The MPL has shown capability of detecting sub-visual cirrus clouds at altitudes greater than 10 km and stratospheric aerosols as high as 15 km. Description of Observational Specifications Wavelength of laser pulse: 523.5 nm Length of laser pulse: ~10 ns = 3 m Range resolution (height interval): 300 m Maximum range for cloud base height: 15 m Maximum profile range: 200 bins * 300 m/bin = 60 km Theory of Operations The principle is straightforward. A short pulse of laser light is transmitted from the telescope. As the pulse travels along, part of it is scattered by molecules, water droplets, or other objects in the atmosphere. The greater the number of scatterers, the greater the part scattered. A small portion of the scattered light is scattered back and detected. The detected signal is stored in bins according to how long it has been since the pulse was transmitted, which is directly related to how far away the backscatter occurred. The collection of bins for each pulse is called a profile. A cloud would be evident as an increase or spike in the back-scattered signal profile, since the water droplets that make up the the cloud will produce a lot of backscatter. Calibration Theory Little calibration is necessary for cloud-base height determination. To fix the distance scale, it is necessary to use a calibrated-pulse generator capable of producing a trigger pulse and a second delayed pulse with an accurately known time lag. The two pulses are used to mimic a transmitted laser pulse and detected backscatter pulse with time delay relating to a simulated distance. Calibration History This information is currently unavailable Current Status and Locations ARM owns five MPL instruments. Three low- resolution MPLs are located at the SGP central facility, at the ARCS-1 on Manus, and temporarily at Pacific Northwest National Laboratory for use at ARCS-3. Two of the five units are high-resolution MPLs (HRs), one of which is at the NSA Barrow site and the other which is being shipped to Nauru for ARCS-2. A significant interest exists in upgrading the MPLs to MPL-HRs, especially the unit at the SGP central facility where comparisons with ceilometers, the Raman lidar, and the MMCR can readily be made. Instrument Mentor notes on data quality control procedures QC frequency: bi-weekly basis QC delay: next week QC type: graphical plots Inputs: raw data Outputs:raw cloud base height (cbh) estimates compared with BLC and high-resolution MPL cbh estimates; processed backscatter profiles Routine data quality monitoring of the MPL at the SGP consists mainly of cross comparison of raw MPL cloud base height (cbh) estimates with those from the BLC and with the high-resolution MPL (MPL-HR) when it is operating there. (Currently, both MPL-HR units owned by ARM are back to the vendor for refocusing.) Comparisons of cbh have been limited by the current ability of the MPL algorithm to generate only one cloud height. Work with Dave Turner (PNNL) and James Campbell (NASA/GFSC) is nearly complete which will allow the MPL to report up to five cloud layers. Besides this rough comparison of cbh, evaluation on a biweekly basis of the processed backscatter profiles produced by NASA GSFC have been important to identify gradual laser degradation in the MPL. The plots, produced weekly, help to discern gradual changes as well as variations with a diurnal cycle. Because the plots are normalized to a uniform color scale, decreases either in the transmitted laser power or in collection-detection efficiency are evident as increased observable scatter in the plots. After the currently known discrepancies between the cbh estimates from the MPL, the BLC, and the Vaisala ceilometers have become understood, development of an automated comparison of reported cloud base heights is recommended. Presently, it would be useful to have flags generated whenever any of the monitored voltages or temperatures drift out of set limits. Perhaps the QC limits already specified in the .b1 level data can be caused to generate e- mail messages or other alerts. Explanation of Flags Applied During Data Ingest Besides actual cloud-base heights, there are two sentinel values present in the data. A cloud-base height of 60 kilometers is reported for a "blocked beam" condition. Literally, this condition means that insufficient backscatter has been at even the nearest bin at 270 meters. This situation can arise from heavy fog or can be caused by water, ice, or debris accumulating on the view port window. A cloud-base height of zero meters is reported for a clear sky. 2.1 Detailed Format Description The ARMCART Belfort Laser Ceilometer data contains eight metadata parameters and eight data parameters and flags. The metadata parameters describe the date, network, station and location at which the data were collected. Data values are collected at the time of observation. All times are UTC. Table 1 below details the data parameters. The data parameters have an associated Quality Control (QC) Flag. A description of the possible QC flag values is listed in Table 2. Table 1 ------- Parameters Units ---------- ----- Date of Observation UTC Time of Observation UTC Network Identifier Abbreviation of platform name Station Identifier Network Dependent Latitude Decimal degrees, South is negative Longitude Decimal degrees, West is negative Station Occurence Unitless Station Elevation Meters Ceiling Height (first layer), QC flag Meters Ceiling Height (second layer), QC flag Meters Ceiling Height (third layer), QC flag Meters Ceiling Height (fourth layer), QC flag Meters Ceiling Height (fifth layer), QC flag Meters Table 2 - Quality Control Flags ------- QC Code Description ------- ----------- U Unchecked G Good M Normally recorded but missing. D Questionable B Unlikely N Not available or Not observed X Glitch E Estimated C Reported precipitation value exceeds 9999.99 millimeters or was negative. T Trace precipitation amount recorded I Derived parameter can not be computed due to insufficient data. 2.2 Data Remarks This dataset contains all cloud level data from the ARMCART Belfort Laser Ceilometer and ARMCART MicroPulse Lidar located at station E13: Lamont Central Facility 1. Through comparisons of the MPL, MPLHR, VCeil25K, and BLC a discrepancy is evident between cloud base heights reported from the BLC and the other instruments. From detailed examination of backscatter profiles, this discrepancy has been identified as stemming from a vertical offset of cloud base heights from the BLC. The cloud base heights reported by the Belfort Laser Ceilometer currently have a vertical offset of undetermined origin which results in a positive bias of approximately 110 meters. The vendor has been consulted and suggests the instrument be sent back for recalibration. The timeframe over which this offset exists has not yet been determined, but it appears that the offset is accompanied by artifacts in the backscatter profile. The mentor is working to quantitatively reproduce the offset from the artifacts in the profile. Once this is done, by processing of the backscatter data, it should be possible to determine the extent to which the offset has changed over time and to correct the existing data sets for this offset. This approach would require that an ingest be written to produce netcdf files of raw blc backscatter profiles. 3.0 Quality Control Processing No additional quality control was performed by UCAR/JOSS on this dataset. 4.0 References ARM, cited 1999: Belfort Laser Ceilometer (BLC) Model 7013C [Available online from http://www.arm.gov/docs/instruments/static/blc.html] ARM, cited 1999: Micropulse Lidar (MPL) [Available online from http://www.arm.gov/docs/instruments/static/mpl.html]