The Meteorological Research Institute (MRI) has established a ground-based radiation observation network with facilities in Tsukuba, Fukuoka, Miyakojima, and Minamitorishima, Japan, in order to assess the effects of aerosols and clouds on the surface radiation budget. One of the key instruments of the MRI ground-based radiation observation sites is the POM-02 sky radiometer manufactured by PREDE Co., Ltd., Japan. In order to maintain accuracy in observation by sky radiometer, the reference sky radiometer is calibrated at Mauna Loa Observatory (MLO) of the National Oceanic and Atmospheric Administration (NOAA). The calibration constant of the reference sky radiometer is determined using the Langley plot technique. The accuracy of the sky radiometer observation at each site is maintained by side-by-side intercomparisons among the reference sky radiometer and the other MRI sky radiometers.

Haze was observed in the Kyushu and Okinawa areas in early August 2020. In Fukuoka, haze was observed from the afternoon of August 2 to midnight on August 5 and from the afternoon of August 6 to the afternoon of August 7. In Miyakojima, haze was observed on the afternoon of August 5. At the same time, a satellite tropospheric monitoring instrument (TROPOMI) observed the diffusion and transport of SO₂ from Nishinoshima in the Kyushu and Okinawa areas. Sky radiometer observations at Miyakojima on August 5 and at Fukuoka on August 6 showed that the daily mean values of aerosol optical depth at 500 nm AOD(500) were 1.400 and 1.597, respectively, and those of single scattering albedo at 500 nm SSA(500) were 0.994 and 0.992. The Ångström exponent α was 1.056 and 1.181, respectively. The volume size distribution retrieved from the sky radiometers showed modes around radii 0.2 μm and 2.0 μm, which are related to sulfate and volcanic ash. These results are consistent with electron micrographs and the elemental analysis results of sampled aerosol particles in Fukuoka.

In the south basin of Lake Biwa, some environmental changes have currently become social issues, such as the occurrence of cyanobacterial bloom, the inflow of muddy or turbid water, and the overgrowth of submerged aquatic vegetation (SAV). Since such changes in ecosystems are highly dependent on meteorological conditions and often occur locally, unexpectedly, or intermittently, they can be difficult to detect by means of the current frequency and effort of water quality monitoring and biological surveys. Consequently, it hinders the progress of analysis of factors that cause the ecosystem changes. In this study, we set a fixed point and time-lapse camera on the roof of a lakeside building and investigated how the camera captured the ecosystem changes during April to December 2018. The camera detected the changes of weather, the occurrence of phytoplankton blooms, muddy or turbid water inflows, and the growth and decline of SAV. We found the peak of Red Ratio (RR) of the RGB was consistent with the muddy or turbid water of the photograph, and Green Ratio (GR) of the RGB tended to be high in response to the occurrence of phytoplankton blooms. In addition, the GR of SAV showed constant, and was significantly lower than that of phytoplankton blooms in the water. Thus, the time-lapse process of muddy water inflow immediately after torrential rain can be captured more clearly by mapping with RR, and the SAV cover can be shown with GR.

Mapping seagrass and seaweed bed habitats is essential for the implementation of sustainable development in coastal waters. In this study, we attempted to create orthomosaic maps from drone images of seagrass/seaweed beds taken of the Takenoura area in Onagawa Bay, Miyagi Prefecture, Japan. We conducted 8 surveys in 2019 and 2020 and successfully constructed a total of five orthomosaic maps. On these maps, seagrass/seaweed beds were either clearly visible (3 surveys) or visible but not clear (2 surveys), based on visual inspection. We failed to construct acceptable maps of 3 surveys, and in one case, could not create even a fraction of a composite map. We found that, to be able to create successful maps, it was necessary to establish a wide enough survey area to include fixed ground features such as land or rock outcrops, and to have sufficient overlap between drone images. In this study, the overlap ratio was set at 70 % on both the front and the side. Our results also suggest that drone surveys be conducted: (1) during low tide so that the distance between the sea surface and the seagrass/seaweed beds (seafloor) is shorter, and (2) under conditions of low turbidity so that the seagrass/seaweed beds can be seen through the sea surface. Attaching a neutral density and circularly polarized luminescence filter was also considered effective to reduce sunlight reflection on the sea surface. Although drone surveys have been reported to be effective in shallow waters in which the seafloor is clearly visible, in our study, orthomosaic maps could be created even in deep waters in which underwater features could be seen. Finally, underwater surveys by divers were conducted for ground truthing. Habitats of both seagrasses (Zostera marina and Zostera caulescens) and seaweeds (Sargassum micracanthum and Sargassum siliquastrum) were confirmed.

Land surface temperature (LST) monitoring is used to detect agro-meteorological events, such as drought and cold weather, to assess crop growth and production. LST products derived from the Second Generation Global Imager (SGLI) onboard the Global Change Observation Mission-Climate (GCOM-C) satellite are provided by the Japan Aerospace Exploration Agency (JAXA). GCOM-C was launched on 23 December 2017, and LST data with 250 m spatial resolution have been available since January 2018. Drought or cold weather is detectable as an anomalous event when LST data for the target period is compared with previous long-term data. However, SGLI/LST data have not been archived over the long term, so a comparison requires data from other sources. On the other hand, MODIS on the Terra satellite has operated for almost 20 years already, and how long continuing useful data can be expected is uncertain. Thus, its data compatibility with other sensors must be assessed to create a comprehensive long-term dataset. The mutual use of SGLI and MODIS LST products would be ideal.

This report compares SGLI/LST (level 3, version 2) images of cropland with the corresponding MODIS/LST (MOD11C1, C6) ones. Comparing a variety of acreage in the world has shown that, for LST above 290 K, SGLI/LST estimates LST at a higher value than MODIS/LST. SGLI/LST data were corrected using a regression model derived from information about the comparability of SGLI/LST and MODIS/LST cropland data to reduce the differences between the data from both sources. This was done on a global scale. Spatially averaged LSTs were also calculated from corrected SGLI/LST and MODIS/LST for six provinces and main croplands in the world. The time-series changes in these data were compared. These products showed a high correlation in time-series changes; the average differences in LSTs for the five provinces ranged from -1.54 to 0.84 K.