The total sky imager (TSI) provides time series of hemispheric sky images during daylight hours and retrievals of fractional sky cover for periods when the solar elevation is greater than 10 degrees. The instrument works by taking an image of the sky with a convex mirror every 30 seconds. The TSI begins capturing images when the sun is 0 degrees above the horizon and begins to process these images into cloud fraction values when the sun is 5 degrees above the horizon. To reduce the saturation of direct sunlight, a black band is placed on the mirror to reduce the sun's direct reflectance. The mirror rotates at the same rate as the sun across the sky. A common problem occurs when the sun position does not match the black band and needs adjusting. The local solar noon image is the best time of day to check the alignment. The noon image is shown with the plots if available. Also, a link to the MPEG movie from the tsimovie datastream is provided on the Diagnostic Plots page if the movie is available. Only the last 3 months of data is available.
The TSI has four datastreams.
tsiskycover – NetCDF data file containing fractional sky cover and sun obscuration by cloud
tsiskyimage – hemispheric sky image every 30 seconds (JPEG)
tsimovie – daily movie of hemispheric sky images (MPEG)
tsicldmask – processed fractional sky cover image every 30 seconds(PNG).
The Yankee Environmental Systems, Inc. TSI-440 FAQ site has some very useful information. Not all of the questions are relevant to the DQ Office so you'll need to sort through them for now.
The TSI has four masks in addition to the cloud masks. First is the mask around the sun (region_sun), which has a 25 degree radius. Next is the horizon mask (region_horizon), which is 40 degrees high by 50 degrees wide. Then there is the 50 degree radius zenith mask (region_zenith). Finally, there is the shadowband mask which covers the shadowband. It is defined as a function of the camera to mirror distance.
This portion of the TSI has two simultaneous analysis regions. The first is the standard 80 degree radius (160 degree field of view) used to determine the "count_opaque" and "count_thin" values. The second is the user defined 50 degree radius (100 degree field of view) used to determine the "region_zenith_count_opaque" and "region_zenith_count_thin" values.
This is the raw image from the TSI data stream.The image with masks.The two rings are an estimate of the data analysis zones. Anything outside of them (i.e. the ground) can reasonably be ignored as it will not likely affect data quality.
The TSI only collects data during the daylight hours, so data during the nighttime hours will always be flagged as Not Available. Sun Strength is usually flagged as Missing.
Plots show the percent opaque and thin and solar elevation throughout the day. There is also a week view of these plots.
Noon image
This plot is an image taken at local noon.
Mask noon image
This plot allows the viewer to better see where the sun is in relation to the shadow band and could possibly determine if the shadow band is misaligned.
List of past problems for this instrument that DO need to be mentioned in your DQA's:
Camera out of focus
This image (4/12/2006 18:30:00) is slightly blurry, noticed by the fuzzy building on the horizon and not so crisp line of the shadowband. See DQPR 1124. The corrected focused image should be more crisp like this one (4/16/2006 18:30:00), although there are some water spots on the mirror.
Shadow Band Alignment
This SGP CF image is correctly demonstrating the normal position of the shadow band at local solar noon. Depending on time of year and location the band could point up or down. The important thing is that it is vertical. Also shown are the bird deterrent spikes to stop birds from pecking at the mirror. This NIM M1 image is showing the shadow band possibly out of alignment. Because the sun is directly overhead it is still properly shading the sun, but the band is not pointing vertically. This particular case of Niamey deployment is under review. The sun is directly over head and may be the reason for the shadow band miss-alignment. Also notice the shadow cast from the camera when the sun is directly overhead.
Pixelated images
Vertical stripes and incorrect colors in TSI imagePercent Opaque and Percent Thin incorrect. Sun/No Sun most likely incorrect.
Water infiltration into lens
Water entered the camera and camera lens. This made imaging the mirror impossible. See DQPR 556
Mirror stopped rotating
Mirror way off of this 5/16/2006 solar noon image. The shadowband should be pointing vertical to cover the sun. The camera has been damaged and is only showing a red image.
The mirror stopped rotating and then allowed the full solar intensity into the camera. This overloaded the camera and damaged the CCD. See DQPR 1157.
Rust forming on the mirror
Rust started to form on the surface of the mirror, and gradually became worse each month. The left -> right pictures are from 7/17, 8/17, 9/17 respectively. See DQPR 1215
Equatorial TSI Shadowband
A slight misalignment of the TSI shadowband for date 3/27/2006 The shadowband migrating from top to bottom of image during transition period. Notice the symmetric shadow of the camera on the mirror as the sun passes directly overhead, and the center of the sun directly in the center of the mirror. Dates 8/10/2006 & 8/25/2006, respectively.
The NIM deployment was near the equator, which resulted in the sun moving from the south side of nadir to the north side during the deployment (and visa versa). This requires the shadowband to change from operating on the bottom of the image to the top. In the transition time the shadowband is capable of shading the sun even if it is not in the normal local solar noon vertical position. Although, this transition period should not last longer than a few weeks to a month. During this period the alignment should be monitored carefully to distinguish a shading problem from normal migration between hemispheres. See DQPR 1608.
Missing bird spike
Notice that the upper right bird spike is missing. While this does not affect data quality, it should definitely be mentioned as it can lead to large obstructions (see below).
Large obstructions
There is a bird on the TSI which is blocking the view of the sky. At the SGP site, bird spikes are placed on the instrument to deter them, but the spikes occasionally break off and need to be replaced. This is a relatively frequent occurrence but should be mentioned in the DQA. Occasionally, a bird will perch higher up. Even though it may be visible, it is not necessarily blocking enough of the sky to be considered an obstruction. In this case, the bird's presence doesn't need to be mentioned. This holds true for other large obstructions such as geckos and bigger bugs. Swarms of smaller bugs may cause a problem if they stick around long enough as well.
Video Splicing
Occasionally, the processes that make the videos will be interrupted. If this occurs, the tasks will be completed when the system is restarted. Subsequently, the video created will likely contain images from the previously interrupted run as well as the ones queued up for that day. This means that part of a previous day's video may be added on to the following day's. This doesn't affect data quality but should be mentioned in the DQA.
Snow Accumulation
When snow accumulates on the mirror, it blocks the view of the sky. Please DQPR if the snow remains on the mirror for an extended period of time.