CoastWatch
Dataset Summary
West Coast Regional Node

Home | Information | Data | El Niño | Software | Sites | Feedback | About Us


Clouds from AVHRR (CLAVR-1) Day-time Cloudmask, Alaskan Region

Program: NOAA Polar-orbiting Operational Environmental Spacecraft
Spacecraft: NOAA-17 and NOAA-18
Sensor: Advanced Very High Resolution Radiometer (AVHRR)
Data Stream: High Resolution Picture Transmission (HRPT)
Primary Geophysical Parameter: Cloud presence
Nominal Accuracy: unknown
Spatial grid: 0.0125 degrees longitude by 0.0125 degrees latitude, geographic
Spatial coverage: Bering Sea, Gulf of Alaska and surrounding areas; 45 to 75 degreess north latitude, 165 degrees east to 125 degrees west longitude
Temporal Coverage: 2006-04-19 to present
FGDC Metadata     Data Quality Act Documentation
Short Description:
The Clouds from AVHRR (CLAVR-1) cloudmask (Stowe, 1999) is used to mask surface temperature data from the Advanced Very High Resolution Radiometer (AVHRR). The cloudmask is offered by CoastWatch so users may examine which tests within the cloudmask indicate cloud presence. The cloudmask should be applied to the unmasked AVHRR All Surface Temperature.
Technical Summary:
In the CLAVR (Clouds from AVHRR) cloud mask algorithm, a series of tests is applied to each pixel in an image. Each test is designed to distinguish between a cloud covered surface and a cloud free surface. Often, a test will target a specific type of cloud. The test can detect that cloud based on its properties and distinguishing characteristics, such as its radiative transmission or reflectance at the wavelengths measured by the AVHRR sensor. These tests are then run in sequence so that all cloud types are detected. The following table summarizes the tests run by the CLAVR algorithm for daytime and nighttime scenarios. Possible return values are shown on the left, along with the flagged test. A zero value is returned when the pixel is determined to be cloud free. These tests are combined into an 8-bit binary number. The first digit (the digit fartherest to the right) represents test #1, the second digit test #2, and so on. This binary number is then converted to an integer value.

8-bit CLAVR ocean cloud mask performed in SST retrieval

Bit #
Daytime
Nighttime
Bit 1
Reflective Gross Cloud Test (RGCT)
Thermal Gross Cloud Test (TGCT)
Bit 2
Reflectance Uniformity Test (RUT)
Thermal Uniformity Test (TUT)
Bit 3
Reflectance Ratio Cloud Test (RRCT)
Uniform Low Stratus Test (ULST)
Bit 4
Channel 3 Albedo Test (C3AT)
Four Minus Five Test (FMFT)
Bit 5
Thermal Uniformity Test (TUT)
Cirrus Test (CIRT)
Bit 6
Four Minus Five Test (FMFT)
Channel 3B Albedo Test
Bit 7
Thermal Gross Cloud Test (TGCT)
Channel 3B Albedo Uniformity Test


Reflective Gross Cloud Test -
This test measures the amount of radiation that is reflected from the earth's surface back to space. Cloud cover has a higher reflectance than a cloud free ocean surface. The test measures AVHRR Channel 2 radiation and compares it with a threshold value. If the value measured is greater than threshold, the pixel is marked as cloudy.

Reflectance Uniformity Test -
The sensor's "picture" of an area will tend to be more uniform in the absence of clouds. Conversely, the presence of clouds can make the radiation measurement more variable. This test compares the variability inside the test area with a threshold value. If the variability measured is greater than the threshold, the pixel is marked as cloudy.

Reflectance Ratio Cloud Test -
This test measures the ratio of Channel 2 reflectance to Channel 1 reflectance. In a cloudy environment, this ratio will fall within a range of values. If the measured ratio is within this range, the pixel is marked as cloudy.

Channel 3 Albedo Test -
This test is similar to the Reflective Gross Cloud Test. It measures reflectance in Channel 3 and detects sparse, transparent clouds. The measured Channel 3 radiation is converted to albedo (the percent of incoming radiation reflected by a body). Clouds have a much higher albedo than the ocean surface. If the value measured is greater than a threshold value, the pixel is marked as cloudy.

Thermal Uniformity Test -
Similar to the Reflectance Uniformity Test, except that this test measures the variability in Channel 4. Cloud free areas will exhibit less variability than cloudy areas in this channel. If the variability is greater than a threshold value, the pixel is marked as cloudy.

Four Minus Five Test -
This test takes the difference between Channel 4 and 5 measurements. Thin cirrus clouds tend to produce larger 4-5 values than would be seen in the absence of clouds. If the measurement is greater than a threshold, the pixel is marked as cloudy.

Thermal Gross Cloud Test -
This test uses a Channel 4 temperature measurement. Because Channel 4 measures the infrared portion, it can be converted to a temperature measurement. Ice particles in clouds can have a much lower temperature than seawater. If the temperature measured is below the threshold (the threshold in this case is the freezing point of seawater) the pixel is marked cloudy.

Uniform Low Stratus Test -
This test is similar in theory to the Four Minus Five Test, except that Channels 3 and 5 are used. This test takes the difference between Channel 3 and Channel 5. Clouds can reduce the Channel 3 measurement significantly. This is especially the case with low clouds, such as stratus clouds. If the 3-5 value is below a threshold, the pixel is marked as cloudy.

Cirrus Test -
This test uses the equation: (T3-T5)/T5, where T3 and T5 are the temperatures resulting from the Channel 3 and Channel 5 measurements, respectively. If this test value lies above a threshold, then the pixel is marked cloudy.

Disclaimer:
Despite our best efforts, incorrect data may often appear within near real time data sets. NOAA CoastWatch accepts no liability for use of these data products. It is recommended that these products NOT be used for navigation.
Acknowledgement:
If this data is used for presentation or publication, please acknowledge the NOAA CoastWatch Program and the NOAA NWS Monterey Regional Forecast Office.
References and suggested citations:
Stowe, L. L., P. A. Davis, and E. P. McClain, 1999. Scientific basis and initial evaluation of the CLAVR-1 global clear/cloud classification algorithm for the advanced very high resolution radiometer. J. Atmos. Oceanic Technol., 16, 656-681.

^ Back to main page Privacy Policy | USDOC | NOAA | NESDIS | CoastWatch