Cloud detection performance of spaceborne visible-to-infrared multispectral imagers

Area of Science:

• Earth Observation
• Atmospheric Science
• Remote Sensing

Background:

• Accurate cloud detection is essential for climate monitoring and weather forecasting.
• Existing spaceborne imagers face challenges in cloud detection efficiency over various surfaces.
• The CLoud and Aerosol Unbiased Decision Intellectual Algorithm (CLAUDIA) aims for unbiased cloud detection.

Purpose of the Study:

• To investigate cloud detection efficiency of current and future spaceborne imagers.
• To evaluate threshold tests for cloud detection over ocean, desert, vegetation, semibare land, and cryosphere.
• To guide CLAUDIA users in optimizing detection thresholds.

Main Methods:

• Utilized the CLoud and Aerosol Unbiased Decision Intellectual Algorithm (CLAUDIA).
• Focused on threshold tests for cloud detection across different ground surfaces.
• Analyzed the performance of various spectral bands, including visible and infrared.

Main Results:

• Fewer spectral bands than Moderate Resolution Imaging Spectroradiometer (MODIS) can lead to cloud detection shifts.
• Infrared bands are crucial; imagers without them show up to 17% cloudy shifts over oceans.
• Specific bands (0.905, 0.935 µm) prevent sun glint misidentification; 0.87, 1.6 µm distinguish clouds from desert; 3.9, 11 µm are critical for snow/ice detection.

Conclusions:

• Imager band selection significantly impacts cloud detection accuracy.
• Infrared and specific visible bands are critical for robust cloud detection across diverse surfaces.
• A proposed "cloud detection imager" with 7-8 bands could enhance efficiency.