The Vaisala Ceilometer (CEIL) is a self-contained, ground-based, active, remote-sensing device designed to measure cloud-base height, vertical visibility, and potential backscatter signals by aerosols. It detects up to three cloud layers simultaneously. Model CL31 has a maximum vertical range of 7700 m. The laser ceilometer transmits near-infrared pulses of light and the receiver detects the light scattered back by clouds and precipitation. For more information see the CEIL Instrument web page.
The CEIL does not “see” higher-level clouds. Thus, cloud base heights that exceed approximately 7 km are not normally reported by the CEIL. The CEIL frequently has “seconds-long” data gaps, much like the MPL. However, the frequency of these data gaps is much less, making data availability checks much easier for this instrument. Generally, a data gap over one minute is a data availability issue.
The primary variable for the CEIL is cloud base height for three possible levels. In the CEIL metrics table, it is common for there to be frequent missing data flags for alt_highest_signal, first_cbh, second_cbh, third_cbh, and vertical visibility. This is OK. So long as the diagnostic plots look correct, the instrument is working correctly. This is occurring because the data file needs to report something, and a 0 could be misinterpreted as a cloud base height. So when the instrument does not report a cloud base height it is filled with the missing value indicator, -9999.
It is normal for altitude of highest signal, first cloud base height, second cloud base height, third cloud base height, and vertical visibility to be frequently flagged as missing.
bl_height_1, bl_height_2, and bl_height_3 are new to the ceil metrics tables and represent the first, second, and third boundary layer height candidates. These come from the ceilpblht datastream, and can also frequently be flagged as missing. The ceilpblht datastream contains planetary boundary heights below 4,000 meters, and includes data from the top of the planetary boundary layer; also known as the depth or height of the mixing layer.
Active Alarms or Warning
At times there may be an Alarm or Warning active. This can be due to a dirty window or another issue with the instrument. If there is a Warning active, the status flag will be yellow. If there is an Alarm active, the status flag will be red. Flags only active for a minute at a time do not need to be mentioned, unless specified otherwise by the instrument mentor, but any enduring Warnings and Alarms need to be included in the DQA.
This is an example of a normal VCEIL backscatter plot on a cloudy day.
PBL Comp
This plot shows the first, second, and third planetary boundary layer heights.
Cloud Base
The first plot shows the detected cloud bases and their heights. The flags are color coded to indicate how many obscurations were detected by the VCEIL. The second plot is the status flag which reports what the instrument is currently reporting about its operation. A Warning level could be an indication of a dirty window or some other possible problem. This may or may not be important and would be additional information if the backscatter plot is showing signs of a dirty window or some other issues. The Alarm level is an indication of a serious problem and should be investigated. If the instrument reports Alarm Level the cloud data will be invalidated. Flags active for only one minute typically do not need to be reported, unless specified by the instrument mentor.
Diagnostic
These plots show different diagnostics for the instrument.
This plot shows a comparison between the VCEIL and the IRT IR Sky Temperature.
TSI Comparison
This plot shows a comparison between the VCEIL and the cloud cover detected by the TSI.
LIDAR Comparison
This plot shows a comparison between the VCEIL and other profiling instruments on site. The cloud bases detected by the VCEIL are indicated on the other instruments' plots with a white dot or line for easier comparison.
Problems that do need to be mentioned in DQAs are mentioned below:
Abnormally low backscatter plot amplitudes
This is a backscatter plot with abnormally low amplitude. Please refer to DQPR 500 for more information.
Consistently high amplitude backscatter
This figure indicates a consistently high amplitude backscatter plot with no cloud detection. This can be caused by a dirty window, but in this case the instrument was not properly switching between low and high amplifier. This plot alone does not indicate a problem, but if other instruments detect clouds and the VCEIL is not able to detect them, something is wrong. See DQPR 501 for more details.
Wavy backsacatter plots
This backscatter shows a strange, wavy-like plot. According to mentor, this problem looks like “electronic ringing.” The “degraded performance can be due to either poor laser operation or degraded detection optics, both will cause similar problems with the amplification.” Note that cloud base heights are detected normally.
Little difference between day & night backscatter
The ceilometer should have a distinct difference in the random noise of the background between day and night. These two plots show the VCEIL degrading to not showing any difference between night/day on Sept. 16, 2005. Also, the background shows a higher amplitude than typical with orange and red background values all day. The first plot shows a normal day (Mostly green, 00:00 ~ 09:00 & 22:00 ~ 24:00) vs normal night (mostly dark blue, 09:00 ~ 22:00) background amplitude. See DQPR 681 or DQPR 1743 or DQPR 1762
Vertical stripes in backscatter
These vertical stripes in the backscatter image were a result of a faulty fiber optic link causing periodic data corruption. See DQPR 682
Low amplitude ripple effect
A ripple effect sometimes shows up in the low amplitude section of the backscatter plot. Sometimes this is caused by a "dirty window", but ofter it is a symptom of a larger issue. Here is the suggestion of the mentor for this case, DQPR 687: The ceilometer has an internal amplifier that it sets to low or high depending on prevalent conditions. Normally, the gain is automatically set high during clear skies. It appears that during these times, a slight degradation in either the transmitter or receiver is being amplified and produces the "ripple" at low altitudes in the backscatter plot. Currently, cloud base height detection is not affected and no action needs to be taken at this time other than to watch for further degradation. For example, keep an eye out for instances of plane-horizontal cloud base with zero variability in height.
Dirty window
The window became dirty on 2/08 (upper left plot) and the background backscatter continued higher than normal throughout the entire day (upper right plot). Cleaning the window on 2/27 brought the background backscatter down to normal levels and fixed the problem (lower left plot). The window could become dirty fast like this example with bird droppings, or may become dirty over time, making the determination more difficult. See DQPR 1701
Dirty Window method II
A second way to find if the window is dirty is through the Status Flag. Normal operation has the flag set to OK. When the window begins to show signs of accumulated dirt the Warning status will begin to be set. The change will typically be gradual and may take days or weeks. Although, once the Status Flag is continuously showing Warn or Alarm it's time to notify the appropriate people. See DQPR 1701, DQPR 1762
Intermittent receiver alarms and discontinuities
The ceilometer was giving intermittent receiver alarms (only availble on instrument computer) and discontinuities in the backscatter plot. The missing data segments may indicate a problem, but the obvious problem is the horizontal "bars" at ~16:30 GMT. These bars plus the missing data required an upgrade to the firmware. See DQPR 1856This examples demonstrates how the VCEIL will not report backscatter profiles if the instrument detects Alarms. This is an indication of a significant issue and is DQPR worthy.
Tilt Angle Flagged
Occasionally the tilt angle will be flagged due to an incorrect positioning of the instrument. While this should be noted in the DQAs, a DQPR is not necessary unless it persists for several days. As long as the data appears to be good it likely is. This is due to the instrument measurements being compensated for the angle.