Aerosol remote sensing studies have been carried out using polar-orbital Earth observation satellites. JAXA has launched Global Change Observation Mission-Climate (GCOM-C)/Second-generation GLobal Imager (SGLI) at the end of 2017, and Greenhouse gases Observing SATellite- 2 (GOSAT-2)/Cloud and Aerosol Imager 2 (CAI-2) in 2018. Moreover, the next-generation geostationary satellite of the Japan Meteorology Agency, Himawari-8/Advanced Himawari Imager (AHI) was launched on October 7, 2014. The AHI has six bands from visible to near-infrared wavelengths and observes the top of atmosphere radiance at the fine resolution of 0.5-2km at every 2.5/10-minutes, which enables the frequent aerosol observation at same ground points. The synergistic uses of these various imaging sensors on both geostationary and polar-orbital satellites are helpful to understand a complete picture of aerosol distribution in the global scale. For this purpose, we developed the common retrieval algorithm of the atmospheric aerosol properties for various satellite sensors and over both land and ocean. The method was applied to Himawari-8/AHI and GCOM-C/SGLI, and the updated retrieval results are presented. The retrieved aerosol properties are validated using ground observation data, such as Aerosol Robotic Network (AERONET) and SKYNET. In addition, the utilization of the aerosol properties forecasted by a global aerosol transport model for the retrieval are discussed.

Retrieving aerosol properties from geostationary Earth orbit(GEO) has been available with visible channels onboard conventional meteorological imager including GOES, METEOSAT, MTSAT and COMS. With the advent of next-generation, meteorological and ocean color imagers onboard GEO platforms, aerosol properties have been provided in unprecendented temporal and spatial resolution. Mulitple channels in visible and near IR from GEO instruments with full disk imaging in 10 minutes made it possible to track fine structures of aerosols in km resolution. These GEO aerosol products retrieved show much higher number of data points than the LEO, with comparable accuracy. These were demonstrated successfully through the extensive field campaigns including DRAGON-NE Asia in 2012, KORUS-AQ in 2016 and EMERGE-Asia in 2018. Due to high demands for data assimilation, these GEO products are now available in near real time. Application of the GEO dataset has been extended to surface PM estimation, air quality forecast, and public health studies. Data merging among different GEO aerosol products at various level of retrieval shows possibility to improve the data quality. With the launch of GOCI-2 with an additional UV channel in 250 m resolution and Geostationary Environment Monitoring Spectrometer(GEMS) together with next generation meteorological imager will provide better capability to better characterize radiative absorptivity and aerosol layer height.

NASA’s dark target (DT) aerosol algorithm currently provides operational retrieval of atmospheric aerosol characteristics from multiple polar orbiting (LEO) satellites. Recently, the DT algorithm was applied successfully to Suomi-NPP VIIRS observations to continue the DT record of aerosol products after the expected retirement of Terra and Aqua in the next five years. Now, we are adapting the DT algorithm to retrieve from next-generation geostationary (GEO) sensor measurements, including the Advanced Himawari Imager (AHI) on Japan’s Himawari-8 (H8) satellite and the Advanced Baseline Imager (ABI) on NOAA’s GOES-East (or GOES-R or GOES-16) and GOES-West (or GOES-S or GOES-17). H8 is a weather geostationary satellite operating since July 2015, and AHI observes the earth-atmosphere system over the Asia-Pacific region at spatial resolutions of 1km or less. Two ABIs on GOES series of satellites provide high temporal resolution observations over the Americas. These new GEO sensors have 16 spectral channels, including 7 bands with similar wavelengths as the MODIS bands used for the DT aerosol retrieval. Most exciting, however, is that both ABI and AHI provide full disk observations every 10-15 minutes and zoom mode observations every 30 seconds to 2.5 minutes. Therefore, the combination of enhanced spectral coverage with improved spatial and temporal resolution of these GEO satellites adds new capabilities for quantifying aerosol loading and other characteristics including aerosol transport and aerosol/cloud diurnal cycles. In this presentation, we will introduce aerosol retrieval results from AHI and ABI using the DT algorithm. We will also present a framework to integrate aerosol data from multiple sensors in LEO and GEO orbits to produce a 30-minute global aerosol data set. Initial examples of global 30-minute AOD product will be shown.

Multi-Angle Implementation of Atmospheric Correction (MAIAC) is a new algorithm that uses time series analysis and processing of groups of pixels for advanced cloud detection and retrieval of aerosol and surface bidirectional reflectance properties. MAIAC was originally developed for MODIS. Recently, we adapted MAIAC to process AHI-8 HIMAWARI measurements. Initial validation shows good quality of cloud detection, aerosol retrieval and atmospheric correction. In this talk, I will provide an overview of MAIAC processing with validation analysis and performance examples.

Geostationary Environment Monitoring Spectrometer (GEMS) is a hyperspectral sensor onboard Geostationary Korea Multi-Purpose Satellite-2B (GEO-KOMPSAT-2B) to be launched in 2020 with the expectation of obtaining tangible evidences to keep track of anthropogenic chemicals. GEMS is designed to provide UV/visible data with improved temporal and spatial resolution as well as higher spectral resolution, though the introduced high quality sensor also requires well-fitted calibration technique. As one of vicarious calibration targets, deep convective cloud (DCC) has distinct properties making it become a favorable target of visible and NIR sensors and considering certain properties are still applicable to UV radiation, here we aim to confirm the viability of DCC calibration technique for GEMS. In the study, Advanced Himawari-8 Imager (AHI) reflectance (R_Band1) and brightness temperature (BT_Band13) were used to detect DCC with the detection thresholds (BT_Band13<205 K, σ_BT<1 K, σ_R<0.03). Ozone Monitoring Instrument (OMI) and Tropospheric Monitoring Instrument (TROPOMI) were also used substituting GEMS and to select DCC pixels from those sensors, collocation with AHI was pre-processed. In the results, enough DCC pixels were detected sufficiently under the GEMS conditions. Though, the left-skewed distribution of DCC reflectance implied DCC detection should be refined with the addition of visible radiation condition to eliminate thick cirrus clouds and cloud edges. Detected DCC spectra of OMI and TROPOMI were highly dependent on the observation geometry such as viewing and solar angle and the effect occurred differently depending on the sensor. They also showed different spectral features caused by the atmospheric effect of ozone (atmospheric attenuation mainly caused by Rayleigh scattering and ozone absorption) when analyzing the magnitude of rotational Raman scattering of each pixels. Considering the results, offsetting those effects needs to be performed hereafter in the study to reduce the effect of external factors to a minimum for inter-comparison between sensors.

In this study, we combined radiances in both the O₂A and B bands measured by the Earth Polychromatic Imaging Camera (EPIC) to determine optical depth and layer height of dust aeorols over ocean surfaces, and biomass-burning aerosols over vegetated land surface. Carried by the Deep Space Climate Observatory (DSCOVR) satellite that orbits around the earth-sun Lagarange-1 point, EPIC observes the entire sun-lit earth disk every 1 to 2 hours, providing earth-reflected radiances in 10 narrow bands. Here we will introduce the remote sensing principals of this algorihtm, demonstrate case studies with validations, and provide an outlook for next steps.

Marine fog is one of the most dangerous weather hazards that threaten coastal and marine areas. However, marine fog monitoring is difficult because of the lack of in-situ observation data. Accordingly, satellite data has been considered a valuable information for monitoring fog above the ocean. In this study, COMS/GOCI marine fog detection algorithm combined with Himawari-8/AHI was developed based on Decision Tree technique. Hourly available and fine visible channel measurement of GOCI and spatial pattern index were mainly used to detect marine fog in the daytime. In addition, cloud top height data from Himawari-8 was used as a supplementary data because visible channel alone is not enough to remove clouds. Decision Tree method was selected for developing our algorithm because the result is easy to interpret marine fog detection process. The resulting algorithm consists of main marine fog detection process and two post-processing including cloud removal and marine fog boundary detection. Post-processing could provide more stable performance than GOCI-only algorithm. Validation was performed for 2107 and 2018 by comparing the detection results with in-situ visibility data and CALIPSO Vertical Feature Mask(VFM) data. Results show that marine fog can be well detected with this algorithm when there is no cloud above the marine fog.

Clouds radiative effects strongly affect the Earth's climate and clouds play a vital role in the energy balance of our planet. As the latter is modulated by the cloud microphysical and optical properties, various remote sensing techniques have been developed to estimate optical properties from satellite and ground observations. Nevertheless, since satellites essentially measure the solar radiation reflected upwards by the clouds while ground instruments the radiation transmitted by the clouds, discrepancies between satellite and surface estimates have been often reported. Atmospheric conditions, resolution, acquisition geometry, and cloud distribution further contribute to this difference. Based on the results gained from an intensive campaign carried out at the SKYNET station of Chiba University (Japan) in November 2018, we focused this presentation on the assessment of satellite estimates of cloud optical depth (COD) retrieved from AHI/Himawari-8 and SGLI/GCOM-C observations. We compared satellite COD with different ground-based COD datasets, namely, COD a zenith, effective COD and COD spatial distribution, retrieved from surface observations of zenith radiance, global radiation, and spatial radiance distribution, respectively. In this way, we attempted to perform an unbiased comparison by explicitly evaluating most of the factors potentially responsible for the discrepancies between satellite and surface estimates.

The Dark Target (DT) aerosol algorithm, originally developed for the MODIS sensors aboard polar-orbiting spacecraft (Terra and Aqua), has been modified and optimized so that it is deriving well-characterized and useful products from multiple sensors. Launched within the past five years, the Advanced Himawari Imager (AHI) on Himawari-8 and the Advanced Baseline Imagers (ABIs) on the GOES16 and GOES17 satellites are the first geostationary instruments to observe with sufficient spectral and spatial resolution to be used as input for a modified DT aerosol algorithm. The modifications necessary to the DT algorithm to be able to process these data include adjusting for shifts in key wavelength bands, mitigating for the missing 0.55 µm, 1.24 µm and/or 1.38 µm channels on the various sensors, adapting for differences in spatial resolution and dealing with the extremely large size of individual disk image files. After these modifications, we have successfully applied the DT algorithm to AHI and ABI inputs, retrieving aerosol optical depth over land and ocean, and aerosol size parameter over ocean. Here we evaluate our products from one month of AHI and ABI data, comparing with ground-truth AERONET observations and collocated DT retrievals from MODIS. These preliminary results are encouraging for the continued development of a DT-geostationary algorithm and eventual creation of a consistent DT aerosol product that unites multiple geostationary and polar orbiting satellites into a single global picture of the comprehensive aerosol system.

Aerosol type classification is critical in both radiative forcing assessment and aerosol remote sensing. However, the major aerosol types in China is still less understood. This work uses AERONET direct sun and inversion measurements from 47 sites in China, with more than 39000 records obtained between April 1998 and January 2017, and apply K-means and SOM clustering to identify major aerosol types. The aerosol optical parameters selected in the classification include optical depth, single scattering albedo, Angstrom Exponent, asymmetry parameter and complex refractive index. In total, we define four aerosol types, namely desert dust, scattering mixed, absorbing mixed and scattering fine mode, whose major composition is inferred to be dust, organic carbon, black carbon and anthropogenic fine particles. We then analyze the spatial and temporal variability and radiative forcing characteristics of the identified types. It is found that dust aerosols mainly occurred in the spring of northwest and north China, scattering aerosols are more common in summer and autumn, absorbing aerosols mostly occur in the winter and spring during heating period. Meanwhile, we also reveal diurnal changes in aerosol types in China. For all types, the average radiative forcing at the bottom (BOA) and at the top (TOA) of the atmosphere is negative, with the BOA forcing significantly stronger than the TOA. Scattering aerosols often have stronger radiative forcing. The desert dust and the absorbing mixed types have stronger radiative forcing efficiency.

Cloud Cover or Cloud fraction (CF) has been recognized as an important parameter to understand the cloud feedback in the context of climate sensitivity. Examining diurnal variability of CF is key to understanding cloud processes. Although there are many studies on diurnal variation of precipitation in the Indian monsoon region, a comprehensive climatology of CF diurnal variability is missing. Since polar orbiting satellites are not suitable for the task, we analyze Indian Space Research Organization’s (ISRO) geostationary satellite Kalpana1 cloud product to address this issue. CF data from Kalpana1 satellite are available every 30-mins at a spatial resolution of 25 km. Here, we have analyzed 10 years (2009-2017) of Kalpana1 data for the summer monsoon (Jun-Sep) season to establish CF climatology at 3 hourly interval. The diurnal amplitude and phase for the monsoon season are examined. Our analysis reveals two distinct modes of CF over Arabian Sea (AS) and Bay of Bengal (BOB). For example, over most part of the AS, CF is maximum at 2330 IST but CF is highest at 1730-2030 IST over the southern part of the AS. In case of BOB, CF peaks at the same time (1730-2030 IST) in the northern part but in the southern part, it peaks at 0530 IST. On the contrary, late afternoon-early evening (1730-2030 IST) peak is observed over the land. The maximum and minimum CF over the AS (0-20N,58-73E), BOB (0-20N,86-94N) and core monsoon region (20-25N,70-88E) are 0.41±0.12 and 0.22±0.10, 0.670.07 and 0.380.07, and 0.620.11 and 0.250.03, respectively. We present the first decadal climatology of CF diurnal variation over the Indian monsoon region using Kalpana1 satellite data. Our results highlight the spatial heterogeneity in CF diurnal variation that the models must represent in order to accurately simulate the monsoon precipitation.

The dark-target (DT) aerosol retrieval algorithm provides products of spectral aerosol optical depth (AOD) from measurements of multi-spectral reflectance. Developed for Moderate-resolution Imaging Spectroradiometer (MODIS), the DT algorithm has been ported to other sensors in Low Earth Orbit (LEO), as well as to airborne sensors. From these observations, we have the basis for an accurate, multi-decade characterization of global AOD. However, these LEO sensors observe only once or twice a day, thus unable to characterize the diurnal cycle. With the advent of 3^rd generation imagers in geostationary (GEO) orbit, we now have capable sensors to retrieve AOD with similar quality throughout the daylight hours. Over the last year, our team has made progress in porting the DT algorithm to Advanced Himawari Imager (AHI) on Japan’s Himawari-8 (H8) satellite as well as to Advanced Baseline Imagers (ABI) on NOAA’s GOES-16 and 17 platforms. With the use of the consistent DT algorithm on all sensors, we are setting up the framework for truly observing aerosol transport and lifecycle as relates to global climate. This presentation acts as a status report for our project, including progress on characterizing retrieval uncertainties, merging LEO and GEO, and creating a product that users can use.

Fog is defined as phenomena with horizontal visibility less than 1km by water droplets and supercooled water in the atmosphere. Because fog causes damages to life, it is important to detect and forecast fog correctly. In this study, fog detection algorithm was developed from GK-2A/AMI data, the Korean 2^nd geostationary satellite launched in December 2018. Visible channels of GK-2A/AMI have from 500m to 1km and infrared channels have 2km with 10minutes time resolution for full disk. The channels used in this study are 0.64, 3.9, 8.6, 10.4, 11.2 and 12.3μm. The surface temperature provided by KMA was used as the ground temperature data to distinguish fog from low cloud. Fog detection algorithm was separately developed based on the usability of satellite channels, day, night, and dawn. It also considered different characteristics of fog on land, coast and sea. Basically, evaluation factors used for fog detection include reflectance, brightness temperature difference(DCD), local standard deviation(NLSD_VIS, LSD_IR1), difference between fog top and ground temperature(ΔFTa or ΔFTs), BT13.2-BT11.2 and BT10.4-BT12.3. The algorithm has been validated using the 280 visibility data of 1 minute period operated by KMA and CALIPSO data operated by NASA. Thresholds were selected using less clouded fog and non-fog cases. In the case of the reflectivity, the background field was produced by 15-day minimum value composite method based on the detection date. In addition, land cover maps and snow cover data were used to reduce false alarmed pixels in desert and snow covered areas where land surface emissivity and reflectivity is similar with fog, threshold values for each evaluation are determined by optimization processes based on the ROC analysis for the selected fog cases. Currently we are performing threshold optimization and will present results of quantitative assessment of fog detection level.

The Advanced Meteorological Imager (AMI) onboard Geokompsat-2A satellite were launched in 2018. Furthermore, the Geostationary Ocean Color Imager-2 (GOCI-2), and the Geostationary Environmental Monitoring Spectrometer (GEMS) onboard Geokompsat-2B satellites will be launched in 2019 in order to monitoring ocean and air quality over Asia. The above three payloads will observe aerosol over Asia in geostationary earth orbit with high spatial and temporal resolution, however the aerosol products will show different spatial and temporal pattern due to the different instrument characteristics, observing wavelengths following with different aerosol algorithms. In this study, aerosol retrieval algorithm for GEMS with synergistic use of aerosol information of AMI are investigated. Low earth orbit satellites instruments are used to test the algorithm as a prototype. First, Total Dust Confidence Index (TDCI) is developed using IR channels to separate dust aerosols, and then applied to improve GEMS aerosol type algorithm. Statistical analysis showed that accuracy for dust aerosols of GEMS are changed from 72% to 91% by using TDCI for aerosol type selection. Second, changed aerosol type are applied to aerosol retrieval algorithm for GEMS. After applying TDCI for dust aerosol selection of GEMS aerosol algorithm, the retrieved results of aerosol optical depth (AOD) and single scattering albedo (SSA) are improved especially for smoke and dust aerosols. These improved results are consistent with algorithm sensitivity test. Finally, AOD products from GEMS and AMI are combined using the maximum likelihood estimate (MLE) method using weights derived from the root mean square error (RMSE) of the original AOD products. The combined AOD products showed increasing spatial coverage compared to the any of the original products, and the accuracy was comparable to the any of the original AOD products.

Fog affects public transportation, ships, and aviation and requires accurate detection and forecasting. To respond to these needs, the Korea Meteorological Administration (KMA) is currently operating about 280 visibility meters to monitor and forecast the fog. The visibility meters has a high observation frequency (1min.), but it has a low spatial representation, missing and abnormal values. Therefore, we developed a low-pass filter type of quality control method (QC) using visibility data and meteorological observations to analyze the detailed characteristics of fog occurrence on the Korean Peninsula. The QC method is consisted of three steps: 1) calculation of the average and standard deviation (SD) using the past 9-minute data (10 minutes). 2) thresholds were set dynamically from 2 SD to 6 SD in order to reflect the average and temporal variation characteristics of the visibility data. 3) the quality of visibility data is checked using the average and threshold values for the calculation of the 10-minute moving averages visibility. For the qualitative and quantitative verification of the QC, the naked-eye observation and the airport weather data were used. The quality test results were visually and statistically evaluated (correlation, Bias, RMSE) with the initial visibility data and the visual data. The visual and quantitative evaluation of the QC results confirmed that the QC was satisfactorily(R: 0.97, RMSE: 1130m). We used weather observation data (e.g., RH, Wind, Pr.) to remove data less than 1km but which is not fog (caused by the other reasons, such as the snow, yellow dust, etc). In this study, the detailed characteristics of fog occurrences on the Korean peninsula will be presented based on the visibility data of recent 3 years (2016-2018).

Global, high-resolution, vertical profile measurements of gaseous pollutants (O₃ and NO₂) and aerosols are identified as NASA priorities and targeted for competition through the Earth Explorer funding line by the 2017 Earth Sciences Decadal Survey. These observables are currently available in the form of tropospheric columns, but quantification of their global health and environmental impacts requires a major advance in measurement technology that allows profiling of their concentrations in the near-surface layer (0–2 km). Improvement in spatial resolution is also required in order to map spatial variability at the neighborhood (intra-urban) scale. HiMAP (High-resolution Imaging Multiple-species Atmospheric Profiler) is an innovative instrument concept crafted to be a low cost, new generation Global Observing System that will enable these advanced measurements within a 6U form factor. It is designed as a push broom, passive remote sensing instrument with two independent modules (UV-Vis: 300–500 nm for O₃, NO₂, HCHO, SO₂ and CHOCHO; NIR: 680–780 nm for aerosols with the oxygen absorption bands). We will present results from recent Observation System Simulation Experiments that estimate the measurement performance, ultimately targeting new capabilities for quantifying gaseous pollutants and aerosols in the near-surface layer over the globe. We will also report on the status and plan of technology advancement for HiMAP.

Aerosols are the tiny particulate matter suspended in the atmosphere and have catastrophic influence on climate. Their spatial and temporal distribution as well as regional and global impact on climate is unpredictable. Aerosols and clouds are the key parameters in climatic processes; hence the understanding of their interaction is extremely essential. In the present work, we analyze the diurnal cycle of aerosol optical depth (AOD_{440 nm)}, evaluate the performance of MODIS Terra-EOS/ Aqua-EOS sensor level 2 C006 AOD retrievals based on DT, DB, and DT/DB combined algorithms by employing 6 – 9 years of AERONET data and carry out long term trend analysis of MODIS AOD over Pune. The diurnal evolution of AERONET measured AOD_{440 nm} at Pune reveals distinct seasonality. During winter and post-monsoon seasons, AOD_{440 nm} appears to be high during forenoon and low during afternoon. While reverse trend is observed during pre-monsoon season. The evaluation/validation analysis depicts that the aerosol scenario over Pune can be well represented by DT, DB products especially DT/DB combined AODs on the basis of linear regression correlation coefficient (R) and its magnitude. Despite an overall MODIS-AERONET AOD agreement, the MODIS sensors tend to underestimate AOD on account of improper assumptions of surface reflectance and the incorrect selection of aerosol types in view of the slope of the regression line being less than 1. AOD occurrence frequency distribution curves at Pune follow normal Gaussian distribution law. It is found that the Gaussian fit is of good quality as illustrated by the coefficient of determination (R²), which takes values in the range 0.97–0.98 for MODIS Terra-EOS and 0.96–0.99 for MODIS Aqua-EOS sensor for the locations under study. Inter-annual trends indicate existence of increasing AOD_{550 nm} trends for Pune (0.0090 yr^–1, R² (COD) = 0.50.