Application of 1-38 μm imagery for thin cirrus detection in daytime imagery collected over land surfaces
Titel:
Application of 1-38 μm imagery for thin cirrus detection in daytime imagery collected over land surfaces
Auteur:
Hutchison, K. D. Choe, N. J.
Verschenen in:
International journal of remote sensing
Paginering:
Jaargang 17 (1996) nr. 17 pagina's 3325-3342
Jaar:
1996-11-01
Inhoud:
While considerable effort has been expended on research into the analysis of optically thin cirrus clouds, the global detection and accurate identification of these clouds remains inadequate, especially in daytime meteorological satellite imagery collected over land surfaces. Recently, 1·38 μm imagery was recommended for the improved detection of these thin cirrus clouds. Since this channel is centred on a strong water vapour absorption band and water vapour is concentrated in the lower atmosphere, incident solar energy in the 1·38 μm spectral band is strongly attenuated once prior to reaching the Earth's surface and a second time after being reflected back toward space under normal atmospheric conditions. Thus, it has been postulated that any energy measured by an airborne (or space-borne) radiometer operating in this spectral band would originate from scattering off of mid-level water and high-level ice clouds, making even thin cirrus readily detectable. While initial results have been encouraging, more quantitative analyses are needed to assess the value of 1·38 μm imagery as a candidate for the next generation of polar-orbiting meteorological satellites. Thus, this project investigates the potential for improved thin cirrus detection in daytime imagery collected over land surfaces using scenes of nearly coincident Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) sensor data, which collects 1·38 μm imagery, and Advanced Very High Resolution Radiometer (AVHRR) imagery collected by National Oceanic and Atmospheric Administration (NOAA) operational meteorological satellites. Results show that the addition of 1·38 μm imagery significantly improves the detection of optically thin cirrus clouds in these daytime data. However, the 1·38 μm imagery does not mask all energy reflected by all surfaces as suggested previously. Thus, the accurate use of 1·38 μm data in an automated cloud classification algorithm requires the development of a detection threshold that varies with surface albedo in the 1·38 μm band and atmospheric water vapour concentration. Meanwhile, it is concluded that imagery collected in the 1·38 μm band significantly improves the detection of optically thin cirrus clouds in daytime imagery collected over land surfaces and would be a valuable addition to future meteorological satellite sensors.
Tijdschrift beschrijving