Smoke detections and visibility estimation using Himawari₈ satellite data over Sumatera and Borneo Island Indonesia

Smoke as the one of weather hazard that contains large pollutant and affect the major live aspects: health, tourism, transportation and climate. Due to its regular appearance in Maritime Continent Indonesian area South East Asia, it is important to assess the satellite remote sensing Himawari₈ data to detect smoke and model the horizontal visibility as the smoke proxy. Using RGB (red, green, blue) combination, maximum likelihood and backward selection of multiple regression were used to detect and to develop the horizontal visibility model. RGB aerosol and RGB day natural color visually sees only the thick smoke with horizontal visibility observe below 1600 m. The best horizontal visibility model with significant level 95% (probability < 0.05) was develop from combination of band 3 (0.64 µm); band 7 (3.9 µm); and band 14 (11.2 µm) with root means square error value is about 404 m and correlation value is about 0.69. © 2018, Korean Spatial Information Society.