Research

Abstract

We are conducting various researches in satellite remote sensing technology, data analysis technology, and applications of satellite remote sensing data for global climate/environmental researches by using the data from various Earth observing satellites.

Data analysis technology

The Earth observations from space had started in 1970's and have been continued by various organizations around the world. Up to now, almost 40-years of data have been accumulated for some of the instruments and are being used for detecting a glimpse of the global changes including sea ice decrease in the Arctic Ocean. Improvements of observing sensors in spatial/temporal resolution and spectral information, an increase of the number of satellites by constellation concept, and development of new observation technologies are leading the significant glowing diversity and volume of the data. To properly extract the information from such huge dataset, we are conducting researches such as in inter-calibration among multiple satellites/sensors and applications of machine learning technique including deep learning to such dataset. Figure below shows an example of rain-area estimation from Himawari-8 infrared brightness temperatures by convolutional neural network trained with ground-based rain radar data (courtesy of KISHIDA Yuko).

The SAR interferometry (or InSAR) is recently attracting attentions for measuring crustal movement by earthquakes and volcanic eruptions and ground subsidence by human activities. We are working on improving various corrections such as for tropospheric delay by using GNSS data and meteorological model output. Figure below shows an example of ground subsidence around Chiba Prefecture in Japan captured by InSAR technique using ALOS/PALSAR data with corrections by GNSS dataset (courtesy of AOTO Ryota).

Remote sensing technology

Developing and improving satellite-borne instruments are one of the essential parts to gain new knowledges of the Earth. We are conducting researches on new sensing technologies required for future missions and sensor calibration methodologies for acquiring high-confidence data. For the future microwave radiometers, large-aperture antennas to improve the spatial resolution and implementation of reduction mechanism for radio-frequency interference (RFI) will be necessary. Our research aims to establish digital-backend microwave radiometers which enable electric beam steering for scanning large antennas and on-board RFI reduction. Figure below-left shows RFI effects in AMSR2 C-band channels over the North America. Red areas indicate the excess brightness temperatures due to RFI. This makes retrieving soil moisture content by using C-band channels difficult. Figure below-right shows NASA's SMAP satellite, which realized the rotating scan system with light-weighted 6m antenna. Further large-aperture antennas may need electric bean steering.

In addition to the space-borne sensors, ground-based instruments also need to be improved. Recently the Global Navigation Satellite System (GNSS) is being widely used for obtaining location information. The system can also be used for atmospheric and oceanic observation such as water vapor and oceanic wind speed by analyzing the delayed or reflected signal. By utilizing the low-cost receiver systems, we are investigating the possibility of constructing dense observation network.

The engineering values or digital numbers obtained by observing instruments are converted to fundamental physical quantities by the calibration process. Although intensive calibration tests are performed during the manufacturing process, the characteristics mostly undergo a change due to the launch vibrations and severe space environment. Particularly for long-term climate variability monitoring, which requires high-precision measurements, the calibration errors and stability have a decisive influence on the mission success. Through the reprocessing of long-term microwave radiometer dataset, we carry out investigations into calibration technology.

Earth environment research

Satellite observations, which have unbiased and periodical coverage of the Earth, have been providing indispensable data for global climate and environment monitoring. With the spatial and temporal analyses of various satellite dataset, we are conducting global environment researches particularly focused on the atmospheric science such as understanding the global water cycle and cloud-precipitation systems. Figures below show the comparison between monthly sea-surface salinity (left) from NASA's Aquarius and flesh water flux (right) from microwave radiometers data (courtesy of Khanpanya Thanarporn). The sea-surface salinity tends to decrease with the increase of flesh water input. The South China Sea exhibits more complex feature due to the water discharge from large rivers.

Figure below indicates the averaged diurnal variation of Himawari-8 infrared brightness temperature during July 10 and August 9, 2018, when we experienced extreme heat wave in Japan (courtesy of Sharon Ludai Anak Sigat). It is clearly seen that diurnal amplitudes [K] are larger in cities and basins and smaller in forested regions, and those over ocean are quite small. Himawari-8 observes over Japan area with 2.5-minutes interval and provides better statistics by data averaging for cloudy areas.

Satellite measurements have been playing essential roles in various areas including daily weather forecast and yellow-dust monitoring. However, such best practices are still limited. This is mainly due to the information mismatch between satellite data and end-users' requirements such as in information quality and spatial/temporal resolution. Further investigations are necessary to resolve such gaps for promoting social implementation of satellite-based applications. Figures below-left shows the photovoltaic unit in Tokiwa campus. Below-right graph indicates time changes of estimated insolation from Himawari-8 data with support-vector regression and electricity generated by the unit (courtesy of YONEMORI Wataru, pyranometer/photovoltaic unit data were provided by the Center for Information Infrastructure). Discussed the application for the health diagnostics of the photovoltaic units without pyranometer measurements.