There are concerns about the lack of forest management in the tsunami-damaged coastal areas affected by the 2011 Tohoku earthquake. In this study, appropriate band combinations and data acquisition times for Sentinel-2 data were examined to extract bamboo and evergreen vegetation from this area. The data obtained from the WorldView-3 and Preiades-1A high resolution optical sensors were also analyzed. The analysis revealed that the classification using short wave infrared (SWIR) demonstrated higher accuracy than the red-edge 3 or near-infrared (NIR) narrow band in Sentinel-2 data analysis. Additionally, the data obtained before the development of new leaves proved useful for extracting bamboo and evergreen forests.

This study aimed to estimate the paddy rice yield using full polarimetric air-borne X-band synthetic aperture radar (SAR) data. We analyzed X-band Pi-SAR2 data of the early maturing stage of rice crops from Osaki City, Miyagi Prefecture, Japan, which were acquired at two different incidence angles, 32° and 53°. Four-component decomposition and eigenvalue-eigenvector decomposition were applied for the data, and full polarimetric parameters were extracted. The correlation of backscattering coefficient values and the full polarimetric parameters with the paddy rice yield were examined. The contribution of volume and surface scattering and scattering entropy computed from the data obtained at incidence angle of 53° were correlated with the paddy rice yield. These parameters were sensitive to panicle biomass and hence can be possible for estimating paddy rice yield.

The windbreak forests called as “Igune” are planted in the surroundings of farmers' dwellings, in several flat agricultural regions, mainly in the Tohoku region of Japan. They bring many benefits to the rural society, but require maintenance. Satellite remote sensing data are useful for the scientific assessment of Igune, because they have a wid geographical distribution. In this study, a method to analyze SPOT 6 data obtained in early May by K-means clustering is proposed. A dataset based on the multispectral image combined with vegetation indices, a dataset based on the multispectral image combined with texture features, and a dataset based on the pan-sharpened image combined with texture features, were analyzed and compared. The optimum number of clusters was determined to be seven by experimental analysis. To extract an Igune with an area of less than 500 m2, pan-sharpened images, combined with efficient texture features, are suitable. Further, the multispectral image, combined with the vegetation indices, is suitable for extracting Igune with areas more than 500 m2, Therefore, SPOT 6 data are expected to be useful for assessing the distribution of Igune.

The possibility of fully polarimetric synthetic aperture radar (SAR) decomposition to detect the temporal change in grassland condition is examined in this study. The study site is located in the coastal area of the Fukushima prefecture, Japan. This area experienced contamination from radioactive materials coming from the nuclear power plant accident caused by the 2011 Tohoku Earthquake. It also includes the controlled access area for residents affected by the accident. ALOS PALSAR and ALOS-2 PALSAR-2 data were analyzed, and artificial grassland areas for the animal industry were extracted. The foreshortening effect should be considered for evaluation of the target in a mountainous area. The decrease of the surface scattering component and increase of the volume scattering component is found in the grassland area in the controlled access area. The change of the scattering component proportion reflects the land-use change on other facilities.

The total area and spatial distribution of paddy fields greatly affect the regional budget of methane (CH4), the second most influential greenhouse gas contribute to global warming after CO2. Therefore, reliable broad-scale paddy field maps are essential for improving the accuracy of regional methane budget estimations. Here, using two types of satellite images, Sentinel-l Synthetic Aperture Radar (SAR) and Sentinel-2 optical images, we developed a broad scale paddy field map for eastern part of Asia in 2018 with a 30 m spatial resolution. The method used in this study combines the robustness of Sentinel-l SAR images to cloud cover with the high accuracy of Sentinel-2 optical images in extracting irrigated paddy fields. By comparing with statistic values of paddy field areas and existing paddy field maps, our paddy field map shows a great potential for reproducing paddy field areas at country scale.

Damage detection due to Typhoon Hagibis was attempted on an agricultural field in Miyagi Prefecture, Japan, using Synthetic Aperture Rada (SAR) data. Sentinel-l data obtained immediately following the heavy rainfall was analyzed. Additionally, the possibility for damage detection by ALOS-2 PALSAR-2 data obtained four days after the rainfall was examined. Image composition of Sentinel-l data obtained directly after the rainfall and the flooding time during paddy rice transplanting is effective in detecting heavily flooded areas caused by dike breach. Flooding because of heavy rain causes a decrease in the backscattering coefficient in both the inside of the field and footpath between the fields. Pixel spacing of PALSAR-2 strip map mode data was 2.5 m and showed the disturbed pattern of the paddy fields caused by the receding floodwater.

We investigate the capability of forest extraction method combining full polarimetric L-band SAR imagery with LiDAR data. The analyzed data are an ALOS2-PALSAR2 image and airborne LiDAR data observed over Osaki city in Japan. In this study, three different types of extraction methods based on an object-oriented classification are compared using a SPOT6 image, the volume scattering component of the PALSAR2 image, and the integrated classification of a volume scattering component of PALSAR2 image and LiDAR data. The accuracy of each method is assessed using the confusion matrix. The results show that method by the integration of volume scattering component image and LiDAR data has a potential to extract forest with high accuracy compared to other two methods.

Marine spatial planning (MSP) has been proposed to facilitate the sustainable use of oceans. To implement MSP in coastal zones, we need to understand the current status of aquaculture, in order to plan for its future. In Japan, coastal aquaculture is based on fishery rights, but it is hard to establish the existing situation as there are so many aquaculture facilities in place. In this study, we used full polarimetric, L-band PALSAR-2 data to identify aquaculture areas. We calculated HH and HV backscattering coefficients, and applied polarimetric decomposition analysis to three-component scattering models. The HV backscattering coefficient successfully discriminated aquaculture areas from the sea surface, while the HH backscattering coefficient was mostly able to discriminate between raft and longline aquaculture areas. The three-component scattering model was then able to discriminate between raft and longline aquaculture types in areas where the HH backscattering coefficient could not. Weather conditions affected the sea surface and therefore influenced the outcomes of our analyses.

Paddy rice field extraction is one of the applications of satellite remote sensing. Full polarimetric synthetic aperture radar is capable of distinguishing between crop types based on the differences in their scattering characteristics. ALOS-2 PALSAR-2 data acquired in early August over a Japanese agricultural area were analyzed to distinguish paddy rice fields from other crops, such as soybean, corn, and vegetables. The scattering characteristics of these crop types were clearly different. The threshold analysis method using polarimetric decomposition parameters examined by a previous study was employed. The results exhibited misclassification of vegetable open culture fields as paddy rice fields. The effect of the variability in the structure of vegetable fields on the scattering mechanism should be considered for improvement of the crop classification method using polarimetric decomposition parameters.

ALOS-2 PALSAR-2 data acquired in full polarimetric mode at the rice heading and well-grown seasons were analyzed to classify agricultural areas according to their use. Eigenvalue-eigenvector and four-component decompositions were used to classify the PALSAR-2 data and discriminate different agricultural parcels. Threshold analysis was applied for the polarimetric decomposition components. Polarimetric decomposition parameters, alpha angle, double bounce scattering component ratio, and surface scattering component ratio, were applied to classify areas according to their use for paddy rice or other crops including soybean. The same threshold value used for analysis using the double bounce scattering component ratio was useful for all PALSAR-2 data analyzed in this study. However, the efficient threshold for alpha angle discrimination was different between the analyzed datasets. Differences in data observation direction, incidence angle, and the growth situation of paddy rice at observation time are conceivable reasons for this difference.

We analyzed airborne L-band Synthetic aperture radar data from the Pi-SAR-L2 platform to detect agricultural fields in semi-mountainous areas of Japan. Full-polarimetric data were acquired during the rice growing season after transplanting. We examined the usefulness of single look complex and ortho-rectified data products by attempting polarimetric decompositions for the single look complex data and examining ortho-rectified images to evaluate image distortions and match these to map coordinates. The three component decomposition images show obvious differences between paddy rice and other fields; the former generally show low backscattering while other crop fields show a larger surface scattering component. Agricultural parcels with an area more than 3,000 m2are clearly recognized on the ortho-rectified image, though narrow agricultural fields in valleys were missed due to the effects of foreshortening.

High-resolution optical satellite images of the reconstructed agricultural fields in Sendai Plain damaged by tsunami caused by the 2011 Great East Japan Earthquake were analyzed to observe variability in crop growth. The main crop of the target area is paddy rice obtained through transplantation, and dry and wet direct seeding. Soybean is the second main crop, with some other vegetables being planted. Normalized difference vegetation index (NDVI) was computed from the images captured in early July in 2014, and early July, late July, and middle August in 2015. Growth variability in the same type of crops was recognized. The differences in the growth statuses resulting from the differences in crop types and cultivation methods appeared on the NDVI images. In a comparison of the three images acquired in 2015, the early July image showed the most obvious variability in crop growth.

Full polarimetric space- and air-borne L-band SAR data acquired on the same day were analyzed. The target areas were agricultural fields, including paddy in the maturation stage, that were severely damaged by a tsunami three years earlier. Lodging of paddy was found in some fields. Three-and four-component decomposition and eigenvalue-eigenvector decomposition analysis were performed. On paddy fields, the dominant component was double-bounce scattering. Paddy fields including lodged paddy showed larger volume scattering ratio than paddy fields without lodged paddy in Pi-SAR-L2 data. On soybean, weed, and bare-soil fields, surface scattering was dominant in a four-component decomposition. Paddy field and other fields were separated by a threshold of alpha angle in the eigenvalue-eigenvector decomposition. Subtle differences were found although PALSAR-2 and Pi-SAR-L2 data behaved almost similarly. L-band SAR is suitable for rice field extraction during this season.

The temporal variations (diurnal and annual) in arboreal (εTree) and bare soil (εSoil) dielectric constants and their correlation with precipitation were examined for several trees in Japan. A significant (1 σ (standard deviation) and 2 σ) εTreeincrease is observed after rainfall at 89.8% and 90.5% probability. However, rainfall does not always induce significant εTreeincreases. Rainfall of more than 5 mm/day can induce 1 σ εTreeTree increase at a 59.6% probability. In order to examine whether the increase in εTree affects the L-band σ0variation in a forest, the four-year temporal variation of the L-band backscattering coefficient (σ0) was estimated from observations by the Advanced Land Observing Satellite Phased Array type L-band Synthetic Aperture Radar. Observed maximum absolute deviations from the mean over the forest area were 1.0 and 1.2 dB for σHH0and σHV0, respectively, and 4.0 and 3.0 dB over open land. σ0and rainfall correlations show that εTreeand σForest0are proportional to precipitation integrated over seven or eight days; εSoiland σOpen land0are proportional to precipitation integrated over three days. This finding indicates that εTreevariations influence σForest areas0. A stronger correlation between σHV0and precipitation is observed in several sites with low σHV0, where less biomass is expected, and several sites with high σHV0, where more biomass is expected. A weaker correlation between σHV0and precipitation is observed for several sites with high σHV0. These differences may be explained by the different contributions of double bounce scattering and potential transpiration, which is a measure of the ability of the atmosphere to remove water from the surface through the processes of transpiration. The two other results were as follows: 1) The functional relation between aboveground biomass and σ0showed dependence on precipitation data, this being an effect connected with seasonal changes of the εTree. This experiment reinforces the fact that the dry season is preferable for retrieval of woody biomass from inversion of the functional dependence of SAR backscatter and for avoiding the influence of rainfall. 2) The complex dielectric constant for a tree trunk, which is measured between 0.2 and 6 GHz, indicates that free water is dominant in the measured tree.

PALSAR full polarimetry data were used to detect thaw/freeze conditions on the ground to try to determine the cause of a spring flood on a northern river in Russia. Full polarimetric data was used to identify the forest area, and the radar backscattering coefficient, σ0HV, was examined. The σ0HV over a frozen forest area was -19.9 dB, while that of a thawing forest area was -11.9 dB, and freeze/thaw events over the forest area were well represented by this parameter. Using this characteristic, together with optical sensor (AVNIR-2) data and microwave radiometer (AMSR-E) data, freeze/thaw conditions were examined for an area including Markha, where a large ice jam flood occurred in 2007, and compared it to 2009, when no flooding occurred. The results indicate that the thawing event proceeded more rapidly between March 29 and April 10, 2007 for the forest, and this rapid thawing may be one of the possible causes for the large ice jam flood in Markha in 2007.

To extract annual changes in paddy field areas, we analyzed COSMO-SkyMed observation data relating to an area damaged by the 2011 Tohoku earthquake. Backscattering coefficient (gamma naught: γ0) values were calculated for images obtained on June 4 2011, June 2 2012, and June 9 2013. Mean backscattering coefficients were calculated for each agricultural parcel using vector boundary data. Areas of relatively low backscattering coefficients correspond with fields that were flooded in relation to rice planting, or fields that have not been repaired. The area of low backscattering coefficients expanded annually between the inland area and the coastal area, showing the progress of paddy field reconstruction. The spatial distribution of COSMO-SkyMed data backscattering coefficient patterns in 2011 were similar to that of TerraSAR-X data obtained 1 day after the COSMO-SkyMed data.

The 2011 off the Pacific coast of Tohoku earthquake was observed using phased array type L-band synthetic aperture radar (PALSAR) and polarimetric and interferometric airborne synthetic aperture radar (PiSAR) full-polarimetric data. Representative polarimetric parameters were calculated from full-polarimetric data for urban areas, where most of the buildings were destroyed by the subsequent tsunami, in order to identify the radar scattering mechanism in these areas. These parameters were compared with the ones observed before the disaster. The full-polarimetric data analysis shows that the affected areas were represented by surface scattering with high entropy, indicating that a complex scattering mechanism with nonreflection symmetry is involved. The coherence between HH and VV and that between RR and LL are the most important factors in distinguishing the disaster areas from the data. Alpha angle and anisotropy are also important factors in this respect; however, anisotropy derived from PiSAR data does not show the difference between areas with collapsed and still-standing buildings. This may be because the azimuth slope angle for the target urban area is different before and after the disaster for both PALSAR and PiSAR data. Owing to the double-bounce scattering from azimuthally rotated targets in the urban areas, the power estimated from the four-component decomposition model is distributed within a wide range not only for double-bounce scattering but also for volume and surface scatterings. Additionally, the model does not show a systematic change between before and after the disaster, and σ0for four polarizations with 30-m resolution does not show a systematic difference.

Using ASTER data, we determined the each specificity of paddy field in the world. Shonai Plain is typical paddy field area in Japan, and has well-developed irrigation and drainage systems. Yangtze River Delta in Middle-East China is the famous paddy fields area in the world, and fishponds are increased recently. Northeast Thai Plain in Thailand and Mekong Delta in Vietnam are famous to produce exporting rice. At the areas, temperature is enough for rice growth, and limitation is water supply. Northeast Thai Plain has severe dry season, and at the season, rice cannot grow. Mekong Delta area is attached South China Sea and the mouth of the Mekong River. At this area, rice grows not only in rainy season but also dry season. Bhutan is small and mountainous country. There are very small size paddy fields in Thimphu area in Bhutan. Sacrament Valley in USA is also very famous commercial rice producing area and the field size is large. At New South Wales in Australia, upland farming fields and paddy fields are mixed, and field size is also large. At east Africa, the consumption of rice is increase, and rice producing is important issue at these countries. The Foot of Mt. Kenya area in Kenya, there is large paddy field area, and well-developed irrigation system like Japan.

Full polarimetric L-band ALOS PALSAR and C-band RADARSAT-2 datasets observing paddy rice field, JAPAN were analyzed by polarimetric decomposition method. From the analysis results by three component decomposition method, remarkable seasonal change is observed by double bounce scattering component ratio on ALOS PALSAR data. RADARSAT-2 data show more than 50% of volume scattering component and it increases with rice growth. Double bounce scattering component reached to 30 40% on PALSAR data, as of August, while it was less than 20% on RADARSAT-2 data. The difference of penetration depth on vegetation area between L-band and C-band may be cause of the difference of the scattering mechanism.

Tropospheric artifact is a serious problem for detecting crustal deformation using differential synthetic aperture radar intereferometry (DInSAR) technique. Advanced Land Observing Satellite (ALOS) has capability for simultaneous observation between synthetic aperture radar, PALSAR and optical sensor, AVNIR-2. We compared ALOS PALSAR interferogram with simultaneously observed AVNIR-2 images. Nominal spatial resolution of AVNIR-2 is sufficient for valuation of PALSAR interferogram. Terra/Aqua MODIS images are compared with PALSAR interferogram and cloud top heights is estimated. Test site is Miyagi prefecture where one of the most active seismic zone in Japanese islands. AVNIR-2 images show the two types of clouds pattern that causes interferometric phase delay. The first is lee wave cloud which shows the presence of gravity waves. The second is thick cloud as cumulus which is formed by condensation of water vapor. An interferogram show combination of phase delays in a couple of data. Simultaneously observed AVNIR-2 image is possible to use for evaluation of PALSAR interferometric fringe patterns.

Multi-polarized space borne SAR data can be used to detect changes in land cover. The ENVISAT ASAR has the ability to acquire images using two polarizations simultaneously. This work discusses the capability of detecting land surface characteristics using ENVISAT ASAR dual polarization data. The applicability of image mode and alternative polarization mode data are investigated, and backscattering coefficients for images with different polarizations are compared. Optical images and GIS data are used as reference data. Two types of land cover in Japan are examined: forest on an active volcano (Miyakejima, a volcanic island that erupted in 2000), and an urban area (central Tokyo). Differences were evident between bare land and forest areas in the HH and HV polarized data sets. City street patterns exhibit differences in backscattering characteristics between like- and crosspolarized SAR images: cross-polarization backscattering is increased for streets that are inclined to the radar irradiation direction.

In repeat-pass SAR interferometry, errors by the tropospheric radar phase delay has been pointed out. We analyze ERS SAR data pairs covering the Kanto Plains, which is a region including Tokyo, Japan. Fringe features that are independent from actual deformations are found from generated interferograms. The fringe patterns are depends on the data; same feature appears on the interferograms generated from the same data. One of the significant patterns is detected in the interferograms from data obtained in August 1999, all of the interferograms by this data contain same fringe feature. A SPOT HRV image acquired 18 minutes after the SAR data reveals patterns of cloud distribution that are similar to the fringe features in the interferograms. It shows a possibility for cloud-induced perturbations in radar signals. Optical sensor imagery with a spatial resolution equivalent to that of the SAR imagery is useful for the evaluation of interferometric fringes.

Investigates the capability of interferometric SAR (InSAR) using C-band (ERS) and L-band (JERS-1) SAR data for land subsidence detection in urban and rural areas. The Kanto Plains and the Saga Plains were selected as the test site. In the Kanto Plains, several continuous land subsidence areas are found both in urban and rural areas, while in the Saga Plains, they are located only in rural areas. We used several data pairs with different time intervals from two to five years from ERS and JERS-1 SAR data, in which observation periods nearly correspond with each other. The land deformation patterns were compared with each other in the data pairs of corresponding observation periods. In the Kanto Plains, the subsidence amounts in urban areas detected by JERS-1 were smaller than those detected by ERS, which might be due to the difference of minimum detectable subsidence. However, JERS-1 could detect major subsidence areas in rural areas, while ERS could hardly detect them. In the Saga Plains, ERS could not detect any subsidence in rural areas, while JERS-1 succeeded in detecting a significant one even by using the data pair with a five-year interval. These results verify that L-band InSAR will be a practical tool for land subsidence monitoring of the Japanese Islands, especially by using ALOS/PALSAR to be launched by Japan in 2004.

We investigate decreasing interferometric correlation of SAR data caused by building damages in urban area. We analyze JERS-1 SAR data pairs straddling the occurrence of the 1995 Hyogoken-nanbu earthquake. The distribution patterns of the pixels which indicated interferometric decorrelation correspond well with the building damaged area reported by the ground survey. The degree of decorrelation shows obvious correlation with the damaged building ratio on the area in each block and measured seismic intensity from strong motion data. The result of this study indicates a fact that the building damage causes the interferometric decorrelation. We also try to extract the urban area damaged by the 2001 Gujrat (western India) earthquake using Radarsat data. For practical use of interferometric decorrelation to detect the urban damage, effect of building type and urban concentration should be examined. These results show the applicability for interferometric analysis of SAR data to urban disaster management.