Improved Cloud Screening Method for the Analysis of Sky Radiometer Measurements and Application to Asian Dust Detection

Abstract

 A cloud screening method employing two successive procedures of variability test and coarse mode test was developed, aiming at better elimination of cloud-contaminated data in the sky radiometer retrievals. The performance of the new cloud screening method was evaluated by examining statistical features with cloud coverage observations and lidar measurements. The variability test appeared to effectively eliminate data contaminated by relatively thick low-level clouds, whereas the coarse mode test appeared to eliminate data likely contaminated by thin cirrus-type clouds. Overall, the new method was considered to improve the current Sky Radiometer Network (SKYNET) data.
 The cloud screening method was then applied to dust detection from sky radiometer measurements. The detection performance was evaluated using surface synoptic observations (SYNOP) dust reports and the yellow sand index from NIES lidar measurements. It was shown that the new method helped to detect dust, effectively eliminating cloud-contaminated signals that were similar to those of the dust.