Evaluating MODIS and MISR aerosol optical depth retrievals over environmentally distinct sites in Pakistan. (February 2019)

• Record Type: Journal Article
• Authors: Ali, Gohar; Bao, Yansong; Boiyo, Richard; Tang, Weiyao; Lu, Qifeng; Min, Jinzhong
• Is Part Of: Journal of atmospheric and solar-terrestrial physics. Volume 183(2019)
• Page Start: 19
• Page End: 35
• Publication Date: 2019-02
• Subjects: AOD -- AERONET -- MODIS -- MISR; Geophysics -- Periodicals; Atmospheric physics -- Periodicals; Géophysique -- Périodiques; Météorologie physique -- Périodiques; Electronic journals; 551.51
• DOI: 10.1016/j.jastp.2018.12.008
• Languages: English
• ISSNs: 1364-6826

Abstract

Abstract: The Moderate resolution Imaging SpectroRadiometer (MODIS) and Multi-angle Imaging SpectroRadiometer (MISR) sensors provide aerosol observations suitable for a wide range of applications. The recently released MODIS Collection 6.1 (C061) includes several improvements compared to the previous products which are expected to minimize uncertainties in aerosol retrievals. Such collection needs to be validated and compared with previous collections form the same or other sensors before being applied for further scientific research on a regional scale. This study evaluates the performance of MODIS Collections 6 (C006) and 6.1 (C061) based on two algorithms: Dark Target (DT) and Deep Blue (DB) and the merged product (DTB) onboard the Terra (MOD04) and Aqua (MYD04) satellites, and the MISR AOD retrievals against ground-based Aerosol Robotic Network (AERONET) over two sites (Lahore and Karachi) in Pakistan during 2007–2014. Results elucidated that C061 MODIS AOD exhibited significant improvement as compared to C006, with the 10 km DT (DB) products generally overestimating (underestimating) AOD relative to AERONET AOD. The MOD04 and MYD04-DT and DB (3 and 10 km) showed comparable performance over the two sites, whereas The DTB was dominated by DT (DB) pixels over Lahore (Karachi). The MISR showed better performance over Karachi with high reflecting surface than over Lahore with dense vegetation cover. The annual cycle of AOD retrieved by the two sensors were consistent with AERONET AOD, with maximum AOD observed during summer months attributed to prevailing climatic conditions. On seasonal basis, the MODIS algorithms exhibited improved performance over Lahore except during summer where DT and DTB showed relatively low performance, attributed to modulations induced by local meteorology to the prevailing surface conditions. However, the sensors exhibited distinct performance over Karachi, where MODIS-DT (10 km) showed close correspondence with AERONET during autumn and winter, whereas MODIS-DT (3 km) exhibited the converse. The MISR performed relatively well during spring over the two stations. The study gives greater insights on the performance of MODIS and MISR and forms the basis for further research on the validation of satellite derived aerosol products over Pakistan. Highlights: Statistical metrics used to assess the reliability of MODIS (Terra and Aqua) and MISR AOD retrieval algorithms against AERONET-AODs over Pakistan. The C061 exhibited significant improvement in AOD retrieval. The MODIS-DT generally overestimated AOD, whereas DB underestimated AODs relative to AERONET AODs. The MODIS-DT performed well in most seasons except during summer. MISR performed well over Karachi as compared to Lahore AERONET site.