Development and Validation of New Glacial Lake Inventory in the Bhutan Himalayas Using ALOS DAICHI

Transcription

1 31 Development and Validation of New Glacial Lake Inventory in the Bhutan Himalayas Using ALOS DAICHI Takeo TADONO 1*, Sachi KAWAMOTO 2, Chiyuki NARAMA 3, Tsutomu YAMANOKUCHI 2, Jinro UKITA 4, Nobuhiro TOMIYAMA 2 and Hironori YABUKI 5 1 Earth Observation Research Center, Japan Aerospace Exploration Agency 2-1-1, Sengen, Tsukuba, Ibaraki , Japan 2 Remote Sensing Technology Center of Japan, Tsukuba, Ibaraki , Japan 3 Research Institute for Humanity and Nature, Kyoto , Japan 4 Faculty of Science, Niigata University, Niigata , Japan 5 Japan Agency for Marine-Earth Science and Technology, Yokosuka, Kanagawa , Japan * tadono.takeo@jaxa.jp Abstract This study aims to develop and validate a new glacial lake inventory in order to gain an understanding of existing conditions. This is being done using optical imageries acquired by the Panchromatic Remote-sensing Instrument for Stereo Mapping (PRISM) and the Advanced Visible and Near Infrared Radiometer type 2 (AVNIR-2) onboard the Advanced Land Observing Satellite (ALOS, nicknamed Daichi ). Glacial lakes can cause outburst floods when natural dams break, and they represent a serious hazard to downstream regions. A major challenge associated with glacial lake outburst floods (GLOFs) is the inability to predict when they will occur or how much damage they will cause. In order to mitigate GLOF hazard risks, a new glacial lake inventory of the Bhutan Himalayas using ALOS imageries is currently being developed, combining PRISM and AVNIR-2 results in imaging with a fine spatial resolution of 2.5 m and multi-spectral precise geolocation. In addition, PRISM can derive precise digital terrain information, i.e., digital surface model (DSM) using its along-track stereo capabilities. We describe the methodology used in the development of the ALOS-based glacial lake inventory, including image processing procedures, glacial lake extraction, accuracy assessments of the generated PRISM DSM for the entire country of Bhutan, and validation of the inventory. Key words: ALOS, AVNIR-2, DSM, glacial lake, GLOF, PRISM 1. Introduction Failure of glacial lakes which are dammed by moraines causes outburst floods and poses a serious hazard to downstream regions. Some researchers believe that global warming has caused the development and expansion of glacial lakes. On the other hand, other researchers state that there is no clear evidence for accelerated expansion from recent warming, especially in the 2000s (e.g., Fujita et al., 2009). Be that as it may, people living in such regions are exposed to the risk of glacial lake outburst floods (GLOFs). This danger exists not only in the Bhutan Himalayan region but also in other cold mountainous regions around the world. To mitigate the damage caused by GLOFs, it is necessary to take a comprehensive approach, including investigations into mass balance, especially melting and shrinking processes of glaciers, evolution processes of glacial lakes, the potential of glacial lakes for outbursts, GLOF trigger mechanisms, and preparations for risk management, such Global Environmental Research 16/2012: printed in Japan as operational monitoring, hazard mapping and early warning systems. An international research project entitled Study on Glacial Lake Outburst Floods in the Bhutan Himalayas has been implemented since 2009 as one of the Science and Technology Research Partnership for Sustainable Development (SATREPS) programs supported by the Japan Science and Technology Agency (JST) and the Japan International Cooperation Agency (JICA). The project aims to achieve comprehensive adaptations for GLOF risks in Bhutan (Fujita et al., 2012). One of the tasks of the project is the development of a new glacial lake inventory using satellite imagery in order to gain an understanding of existing conditions and the expansion history of glacial lakes in Bhutan. Recently, various types of satellite data with different resolutions, recurrences, numbers of spectral bands, functional capabilities and costs have become available. The accuracy of satellite data has also significantly improved. They are effective in monitoring surface 2012 AIRIES

2 32 T. TADONO et al. conditions, including in remote regions where most of the mountainous glaciers and glacial lakes in the world are located. In addition, data from a different set of satellites make it possible to identify temporal changes and put them into a historical context. Although satellite data are definitely not a panacea for monitoring everything, they have become one of the most efficient and important tools. The objective of this study is to develop a new glacial lake inventory for the Bhutan Himalayan region by utilizing the Panchromatic Remote-sensing Instrument for Stereo Mapping (PRISM) and the Advanced Visible and Near Infrared Radiometer type-2 (AVNIR-2) onboard the Advanced Land Observing Satellite (ALOS, nicknamed Daichi ), which operated from January 2006 to April The latest glacial lake inventory for the Bhutan Himalayas was published in 2001 by the International Centre for Integrated Mountain Development (ICIMOD) in Nepal (Mool et al., 2001). This inventory was based on Landsat-5 TM data (30 m resolution) in 1999, and provided the information for all glacial lakes in the Bhutan Himalayas. It was reported that expansion of some glacial lakes was noted; however, changes in the conditions of glacial lakes in the Bhutan Himalayas since 2000 were not incorporated. This study explains the procedure for establishing the newly developed inventory in the following steps: image processing, lake extraction, and screening of extracted lakes to identify glacial lakes. The digital surface models (DSMs) were generated by PRISM for the entire country of Bhutan. The generated DSMs along with the ALOS-based inventory were validated by comparing them to the existing digital elevation model (DEM) and groundbased global positioning system (GPS) measurements. 2. Data In this study, our primary focus was on the data processing carried out by PRISM and AVNIR-2 onboard ALOS, which was launched on January 24, 2006, and continued to operate well even after the mission target life of five years (Shimada et al., 2010). However, its operation moved to the low load mode (LLM) on April 22, 2011 due to a power generation anomaly. This was subsequently followed by an official termination of its mission on May 12, Notwithstanding, approximately 6,500,000 scenes of archived data covering the entire globe are available to users. ALOS had three mission instruments: an L-band Synthetic Aperture Radar called PALSAR and two optical sensors. PRISM consisted of three radiometers: nadir-, forward-, and backward-looking, and had a 2.5 m spatial resolution with a 35 km swath width in triplet mode. It performed along-track stereo observation to generate precise DSM. AVNIR-2 had four radiometric bands ranging from visible to near infrared, and a 10 m resolution with a 70 km swath width. They were both able to observe the same area simultaneously. Other satellite imageries such as KH-9/Hexagon, SPOT-1 and 2, Landsat, JERS-1/OPS, and Terra/ASTER were used primarily to investigate the expansion history of the glacial lakes in the project. These results will be included in the final version of the glacial lake inventory. In cases where the above-mentioned instruments had stereo capability, we attempted to extract the terrain height information, i.e., DSM or DEM information, because such information is beneficial for evaluating the potential of GLOFs if the data were acquired over the lakes before a previous GLOF event. We conducted field surveys in 2010 and 2011 in order to collect ground control point (GCP) data, take continuous measurements while trekking and make lake size measurements from portable GPS receivers. All groundbased GPS measurements were post-processed using base-station GPS data. Some GCPs were used in processing the DSM and the ortho-rectified images (ORIs). The measured glacial lake areas were used to validate the glacial lake inventory. 3. Methodology and Results First, we established a procedure for the development of the glacial lake inventory by using satellite imageries from image data processing that included ORI generation for all satellite data used in our analyses, DSM generation when possible, and the extraction of glacial lakes. It was necessary to correct the geometry of all satellite data in order to compare them as they were acquired using different satellite instruments with varying spatial resolutions, because one of the purposes was to investigate the glacial lake expansion history. In this study, we mainly describe the data processing procedure using PRISM and AVNIR-2. Figure 1 shows the data processing flowchart of PRISM and AVNIR-2 aboard ALOS for developing the glacial lake inventory in this study. Ukita et al. (2011) and Tadono et al. (2011) reported on progress in the processing and preliminary results of the glacial lake inventory. The image data processing starts from ORI generation for both PRISM and AVNIR-2 and DSM generation for PRISM, for which our software, called the DSM and Ortho-image Generation Software for ALOS PRISM (DOGS-AP), is used (Takaku & Tadono, 2009a). DOGS-AP has the functions of data processing and Fig. 1 The data processing flowchart of PRISM and AVNIR-2 for the development of the glacial lake inventory.

3 The ALOS-based glacial Lake Inventory Development in the Bhutan Himalayas instrument calibration. Since the launch of ALOS, we have carried out instrument calibrations of PRISM and AVNIR-2 to improve the absolute accuracy of the standard products (Tadono et al., 2009). These results are also reflected in the DOGS-AP software. As a result, GCPs are not essential for PRISM. We validated the generated PRISM DSMs at several test sites with various surface conditions (Takaku & Tadono, 2009b, 2010). DSMs from the airborne Lidar and GCPs were used as reference data. The results showed that the height accuracies of PRISM DSMs fell in the range of 2.94 m to 7.15 m (root mean square error, RMSE), ~3.44 m (bias) and ~6.44 m (standard deviation, STDEV). In addition, we concluded in previous studies that the bias error correction has only to consider the height of the DSM even if GCPs were not used in processing them (Takaku & Tadono, 2011). The ORI of PRISM conducts simultaneous processing with DSM generation. To obtain the above-mentioned height accuracy from the generated PRISM DSM, ORI should also have precise planimetric accuracy, except for low correlation regions, e.g., clouds, oceans and saturated pixels. These pixels were masked out before processing. We adopted a filter in the software to derive height information in such masked areas from surrounding regions. PRISM DSM was generally processed with 0.3 arcsec (approximately 10 m) pixel spacing. ORI generation was also carried out for AVNIR-2 using DOGS-AP with the existing DEM. The DEM created by the Shuttle Radar Topography Mission (SRTM) was used for Bhutan. The absolute geometric accuracy of AVNIR-2 was not as good as that of PRISM due to hardware limitations (Tadono et al., 2009). 33 Therefore, GCPs were necessary for ORI processing. These were collected from PRISM ORI and DSM if there were no ground-based data. The AVNIR-2 images used here were acquired at zero degrees pointing angle, i.e., the nadir observation and the geolocation accuracy of the ORI should therefore be basically within one pixel. Pan-sharpened images were processed using both ORI images from PRISM and AVNIR-2 to obtain 2.5 m resolution color composite images. This is crucial for identifying and interpreting detailed ground features as required in making the glacial lake inventory. The commercial software (ERDAS Imagine) was used in this processing step. We tested a class of pan-sharpening algorithms. Six methods were first selected from among the available algorithms in the software: principal component analysis, multiplicative, the Brovey transformation, modified IHS, high-pass filtering, and subtractive resolution merge (ERDAS Inc., 2008). The subtractive method was finally selected for the image quality criterion, color balance, multiband usage and processing time. Image tie points were collected between PRISM ORI as a master image and AVNIR-2 ORI as a slave image, and then the software automatically generated the pansharpened image. Figure 2 shows a pan-sharpened mosaic image for the Bhutan Himalayas as of July 2011 with bands 3, 2 and 1 assigned as red, green, and blue, respectively, and boundaries of river basins overlaid with a red line. The color dots show extracted glacial lakes. The image covers almost the entire country of Bhutan, except for some cloudy areas. Hence, it can be used not only as a base image for the ALOS period but also as a reference to other satellite imageries thanks to its fine resolution and precise geolocation. Fig. 2 Pan-sharpened mosaic image by PRISM and AVNIR-2 in the Bhutan Himalayas as of July 2011 (red, green, blue = Band 3, 2, 1 as true color composite). The red lines show the boundaries of the river basin and the colored dots show glacial lakes extracted in this study.

4 34 T. TADONO et al. DSM mosaic processing was also carried out to create another base image of Bhutan. There are three steps involved in the processing: stacking, bias correction, and mosaicing. As mentioned above, the planimetric accuracy of PRISM ORI and DSM is basically high even if processing is done without GCP. However, in this case a bias height error may not be so negligible. Hence, DSM stacking was done simply as a means of averaging the heights of all available scene-frame DSMs except for the masked areas. The bias offset of the DSM without GCP was calculated by comparing height differences in overlapped areas with neighboring scene-frame DSMs, which were processed with GCPs. Figure 3 shows the mosaic image of PRISM DSM in the Bhutan Himalayas as of July The gray-scale corresponds to ellipsoid height, which is defined as Geodetic Reference System 1980 (GRS 80) by the 1997 International Terrestrial Reference Frame (ITRF 97) coordination, and black indicates areas masked due to clouds. It has 0.3 arcsec (~10 m) pixel spacing. Figure 4 gives a visual comparison between PRISM DSM with 0.3 arcsec and SRTM with 3.0 arcsec (SRTM-3) as existing global DEMs. Details of the terrain features can clearly be seen in PRISM DSM. Before identifying glacial lakes in the image, all lakes were extracted as much as possible so as to keep the quality consistent. Pan-sharpened images from PRISM and AVNIR-2 were basically used as the base images for extracting lakes. If an area of the pan-sharpened image was not available due to cloud cover, AVNIR-2 was used instead to extract lakes. We tested automatic and semiautomatic methods of extracting water bodies from the pan-sharpened images. It turned out that these methods were not feasible due to unstable effects caused by terrain, shadows, clouds, frozen water and snow cover. Instead manual digitizing was used to extract water bodies and they were archieved as shape files. This was done using commercial software (ERDAS Imagine and ArcGIS). When a lake was clearly visible and distinguishable from the background, for example, when no ice was in lake, its identification was rather straightforward. In order to ensure that an object was a real lake, we consulted terrain features from the PRISM DSM, e.g., contour lines, Fig. 3 PRISM DSM mosaic in the Bhutan Himalayas as of July Black indicates areas masked due to clouds, and red lines show the boundaries of the river basin. The yellow square represents the magnified area depicted in Fig. 5. PRISM DSM SRTM-3 Fig. 4 Visual comparison between PRISM DSM (left) and SRTM-3 (right) in the same region. The red on the right indicates void data areas in SRTM-3.

5 The ALOS-based glacial Lake Inventory Development in the Bhutan Himalayas 35 terrain slopes, elevation profiles and bird s eye views. This analysis provided additional terrain information with regard to the location of ridges and divides, and to a large extent eliminated false identification in shadow zones. Finally, glacial lakes were screened from the extracted lakes based on the following definition: a glacial lake is a body of water lying between the terminus of the mother glacier and the Little Ice Age moraine that was identified as a sparsely vegetated moraine surrounding the lake (Iwata et al., 2002). Lakes located within 2 km of the Little Ice Age moraine down-valley were also included so as to take into account possible flooding events with multiple lakes being involved. In addition, supraglacial lakes on debris-covered glaciers were included. We also set 0.01 km 2 as the minimum lake size in consideration of the fact that small lakes represent less GLOF risk. Note that this is the definition which we adopted for our inventory. In fact it is different from the one for the existing glacial lake inventory by ICIMOD (Mool et al., 2001). The latter adopts the definition that isolated lakes above 3,500 m a.s.l. are listed as the remnants of glacial lakes left over when the glaciers retreated. The blue, yellow, orange, green, red, purple, pink, light-green and sky-blue dots in Figure 2 represent extracted ALOS-based glacial lakes across the Bhutan Himalayan region as of July Those colors were assigned for different river basins because our inventory was also organized into nine separate river basins in Bhutan: the Amo Chu, Wang Chu, Mo Chu, Pho Chu, Mangde Chu, Chamkhar Chu, Kuri Chu and Dangme Chu ( Chu means river) and the Tibet side. The total number of extracted lakes was 733. This figure is not final as another round of quality control is currently in order. The parameters of our inventory primarily followed the ICIMOD convention, including ID, coordinates of latitude and longitude of the lake center, area, elevation, orientation, width and length of the lake, along with auxiliary information such as type of lake (e.g., moraine-dammed lake, ice-dammed lake, supraglacial lake). Each lake has its own ID based on the name of the river basin and the latitude and longitude at the center of the lake. In addition, we provide a cross-reference with respect to the ICIMOD code whenever available. Table 1 shows the contents of the ALOS-based glacial lake inventory for the Mangde Chu basin as an example, where 39 glacial lakes are listed. The inventory consists of the shape files shown in Fig. 2, and the parameter shown in Table Validations In order to validate the generated PRISM DSM, the ALOS-based glacial lake inventory was compared with reference data provided by ground-based GPS measurements and an existing DEM. First, the generated PRISM DSM mosaic was compared to ground-based GPS measurements consisting of point measurements taken as check points (CPs) in 2010 and 2011, and continuous GPS measurements taken while trekking in These measurements include some inaccurate data, which may be caused by shadows from forests and steep terrain, as well as conditions such as dilution of precision (DOP). Therefore, a set of measurement points within +/ 30 m of errors with respect to the PRISM DSM were selected. This set included a total of 3,268 GPS points consisting of the 3,198 continuously measured points in 2010, 59 CPs in 2010, and 11 CPs in Figure 5 shows a magnified DSM mosaic of the area indicated by the yellow square in Fig. 3. The locations of the continuous measurements are given by yellow dots, CPs in 2010 by red dots, and CPs in 2011 by blue dots. Figure 6 and the upper row of Table 2 summarize the validation results. They confirm that the height accuracy of the PRISM DSM mosaic was 2.3 m of bias error, 7.8 m of STDEV, and 8.2 m of RMSE. These results are relatively consistent with our previous validation results as explained in section 3 (Takaku & Tadono, 2009b, 2010). Another validation effort for the DSM mosaic was carried out using SRTM-3 as the existing DEM. This was done by simply overlaying both sets of data and calculating height differences as PRISM DSM minus SRTM-3 DEM. Figure 7 shows this height difference image between the PRISM DSM and the SRTM-3 for the entire state of Bhutan. White represents areas in which the height difference is around 0 m, while red and blue graduations correspond to areas with around +/ 30 m differences. Note that black indicates areas masked in the PRISM DSM, while green represents void data areas in SRTM-3, which in many areas still remain void due to steepness and other features of glacial terrain. Large errors shown by blues and reds in Fig. 7 are seen around the north central section of the Bhutan Himalayas, which is the highest and most mountainous region of Bhutan, covered with glaciers and snow. These height errors may have occurred in the PRISM DSM due to a lack of image-matching accuracy. However, regions of misinterpolation in SRTM-3 in northeastern Bhutan were not shaded in green as void data. Figure 8 is a histogram of the height difference, and the lower row of Table 2 summarizes the statistical results of the validation. More than seven hundred million DSM heights were compared in this validation, and the error clearly displayed characteristics of Gaussian distribution. The height accuracy of the PRISM DSM mosaic compared with SRTM-3 was 0.4 m of bias and 20.7 m of STDEV and RMSE. In these DSM validation exercises, it is difficult to be certain of the accuracy of the PRISM DSM mosaic because both sets of reference data still remain uncertain. The GPS measurements are limited in terms of area coverage and accuracy, while SRTM-3 has void and interpolation regions likely to have low accuracy. Nonetheless, we can say that the height accuracy of the PRISM DSM is around 8 to 20 m in RMSE. In addition, it has a fine resolution with less than 20 m planimetric accuracy and can provide detailed terrain features, as shown in Fig. 4.

6 36 T. TADONO et al. Table 1 Contents of the ALOS-based glacial lake inventory for the Mangde Chu basin.

7 The ALOS-based glacial Lake Inventory Development in the Bhutan Himalayas 37 Fig. 5 Magnified PRISM DSM mosaic in the northwestern region of the Bhutan Himalayas, and location of GPS measurements points (yellow: continuous measurement in 2010; red: CPs in 2010; and blue: CPs in 2011). Fig. 6 Comparison of height between ground-based GPS measurements and PRISM DSM (red: continuous measurement in 2010; green: CPs in 2010; and blue: CPs in 2011). Table 2 Summary of validation results for the PRISM DSM mosaic in the Bhutan Himalayan region. Reference data Ground-based GPS Total number Bias error (m) Standard deviation (m) RMSE (m) 3, SRTM-3 738,107,

8 38 T. TADONO et al. Fig. 7 Height difference image (i.e., PRISM DSM minus SRTM-3) in the Bhutan Himalayas. Black indicates areas masked in the PRISM DSM, while the green represents the void data areas in SRTM-3. Fig. 8 Histogram of height differences between PRISM DSM and SRTM-3. Second, the ALOS-based glacial lake inventory was validated by comparison with ground-based GPS measurements. Previous results on glacial lake inventories have been reported by Ukita et al. (2011) and Tadono et al. (2011). RMSE geolocation accuracy of 11.7 m and bias of 9.5 m were confirmed at the Metatshota Glacial Lake in the Mangde Chu basin. This large bias error included a gap between the actual shoreline and that measured by walking with the GPS receiver around the lake. We also pointed out large location offsets between our inventory and that of the ICIMOD (Mool et al., 2001). Thus, a part of the difference in the lake size can be attributed to the different observation times. However, this discrepancy appeared at other lakes, and the differences were not consistent. In order to look into this discrepancy, we compared the inventory to ground-based GPS measurements taken in 2011 at two other lakes in the Wang Chu basin, located in western Bhutan. Figure 9 shows a comparison of the polygons between the inventory (yellow) and the ground-based GPS measurements at those two lakes (green) that were overlaid onto the pan-sharpened image. It can be clearly seen that the inventory more or less matches the ground measurements for the westernmost lake. On the other hand, the easternmost lake, called Karma Glacial Lake, has almost the same location between the inventory and ground measurements. Yet the size of lake in the inventory is larger than that of the ground measurements. This size difference is attributable to seasonal changes in water depth, a phenomenon confirmed at Karma Lake during a survey in Seasonal changes in the depth of water in glacial lakes, especially supraglacial lakes, remain a complicated scientific research topic, and we will not attempt to discuss the details in this paper. However, more careful considerations will be necessary when analyzing the expansion history of supraglacial lakes in the future. 5. Conclusions In this study, we have described and validated the development of a new glacial lake inventory for the Bhutan Himalayan region using PRISM and AVNIR-2 onboard the ALOS satellite. We first processed orthorectified pan-sharpened images as well as PRISM DSMs for the entire country of Bhutan. These images are actually used not only for producing the glacial lake inventory but also for other applications, such as providing base data, as there were no digital maps or terrain models with fine enough spatial resolutions. The height accuracy of the DSM mosaic was confirmed to be in the range of 8 m to 20 m RMSE using ground-based GPS

9 The ALOS-based glacial Lake Inventory Development in the Bhutan Himalayas 39 Fig. 9 Comparison of polygons between the ALOS-based inventory (yellow) and the ground-based GPS measurements (green) at two lakes in The background is the ALOS pan-sharpened image (red, green, blue = Band 3, 2, 1). measurements and SRTM-3. Based on these base data, the glacial lake inventory is in ongoing development, and a total of 733 lakes have so far been extracted for the Bhutan Himalayan region. The geolocation accuracy of the inventory was confirmed at two glacial lakes by comparing the inventory data with actual ground measurements taken in Closer attention may be needed for seasonal changes in water depth when considering the supraglacial lake expansion history in the future. The complete version of the ALOS-based glacial lake inventory will be available by March 2012 as planned and open to the public via the web site at < jp/alos/en/bhutan_gli/index.htm>. Acknowledgements We would like to thank Mr. Junichi Takaku of the Remote Sensing Technology Center (RESTEC) of Japan for assistance with the PRISM DSM generation. We would also like to thank Prof. Nozomu Naito, Prof. Takanobu Sawagaki, Dr. Satoru Yamaguchi, Prof. Jiro Komori, and Prof. Koji Fujita for collecting and processing ground-based GPS data in 2010 and This study was carried out under the project entitled Study on Glacial Lake Outburst Floods in the Bhutan Himalayas supported by the Japan Science and Technology Agency (JST) and the Japan International Cooperation Agency (JICA), under the Science and Technology Research Partnership for Sustainable Development (SATREPS). We would also like to gratefully acknowledge all our colleagues of the project for collaborating with us. References ERDAS Incorporated (2008) ERDAS Imagine Field Guide. Fujita, K., A. Sakai, T. Nuimura, S. Yamaguchi and R.R. Sharma (2009) Recent changes in Imja Glacial Lake and its damming moraine in the Nepal Himalaya revealed by in-situ surveys and multi-temporal ASTER imagery. Environmental Research Letters, 4: Fujita, K., K. Nishimura, J. Komori, S. Iwata, J. Ukita, T. Tadono, and T. Koike (2012) Outline of research project on glacial lake outburst floods in the Bhutan Himalayas. Global Environmental Research, 16: Iwata, S., Y. Ageta, N. Naito, A. Sakai and C. Narama, Karma (2002) Glacial lakes and their outburst flood assessment in the Bhutan Himalaya. Global Environmental Research, 6: Mool, P.K., D. Wangda, S.R. Bajracharya, K. Kuzang, D.R. Gurung and S.P. Joshi (2001) Inventory of Glaciers, Glacial Lakes and Glacial Lake Outburst Floods: Monitoring and Early Warning Systems in the Hindu Kush-Himalayan Region, Bhutan, International Centre for Integrated Mountain Development, Kathmandu. Shimada, M., T. Tadono and A. Rosenqvist (2010) Advanced Land Observing Satellite (ALOS) and monitoring global environmental change, Proceedings of the IEEE, 98: Tadono, T., M. Shimada, H. Murakami and J. Takaku (2009) Calibration of PRISM and AVNIR-2 Onboard ALOS Daichi. Transactions on Geoscience and Remote Sensing, The Institute of Electronics Engineers, Inc., 47: Tadono, T., M. Shimada, T. Yamanokuchi, J. Ukita, C. Narama, N. Tomiyama, S. Kawamoto, K. Fujita and K. Nishimura (2011) Development of glacial lake inventory in Bhutan using ALOS Daichi, Proceedings of International Geoscience and Remote Sensing Symposium (IGARSS) 2011, The Institute of Electronics Engineers, Inc., CD-ROM. Takaku, J. and T. Tadono (2009a) PRISM On-Orbit geometric calibration and DSM performance. Transactions on Geoscience and Remote Sensing, The Institute of Electronics Engineers, Inc., 47: Takaku, J. and T. Tadono (2009b) High resolution DSM generation from ALOS PRISM Status updates on over three year

10 40 T. TADONO et al. operations. Proceedings of International Geoscience and Remote Sensing Symposium (IGARSS) 2009, IEEE, CD-ROM. Takaku, J. and T. Tadono (2010) High resolution DSM generation from ALOS PRISM Processing status and influence of attitude fluctuation, Proceedings of International Geoscience and Remote Sensing Symposium (IGARSS) 2010, IEEE, CD-ROM. Takaku, J. and T. Tadono (2011) High resolution DSM generation from ALOS PRISM archiving and mosaicing, Proceedings of International Geoscience and Remote Sensing Symposium (IGARSS) 2011, IEEE, CD-ROM. Ukita, J., C. Narama, T. Tadono, T. Yamanokuchi, N. Tomiyama, S. Kawamoto, C. Abe, T. Uda, H. Yabuki, K. Fujita and K. Nishimura (2011) Glacial lake inventory of Bhutan using ALOS data Part I: methods and preliminary results. Annals of Glaciology, International Glaciological Society, 52(58): Tsutomu YAMANOKUCHI Tsutomu YAMANOKUCHI joined the Remote Sensing Technology Center of Japan (RESTEC) in He is currently in charge of cryospheric applications using ALOS data mainly at mountain glaciers and in Antarctica. He also dedicates efforts to capacity building of satellite data utilization through hands-on training to users in developing countries. He received his B.S. degree in Astrophysics from Nagoya University, Aichi Pref., Japan in 1993, M.S. degree in Oceanography from Kyushu University, Fukuoka Pref., Japan in 1995, and a Ph.D. in Science from the Graduate University for Advanced Studies. Jinro UKITA Takeo TADONO Takeo TADONO is an Associate Senior Researcher at the Earth Observation Research Center (EORC), Japan Aerospace Exploration Agency (JAXA), Tsukuba, Ibaraki Pref., Japan, and is in charge of calibrating and validating optical instruments on the Advanced Land Observing Satellite (ALOS) and ALOS follow-on missions. He received his B.E. degree in 1993, M.E. degree in 1995, and Ph.D. in 1998 in Civil Engineering from Nagaoka University of Technology, Niigata Pref., Japan. His research interests are calibration and validation of high-resolution optical instruments and the development of algorithms for retrieving geophysical and hydrological parameters from remote sensing data. Sachi KAWAMOTO Sachi KAWAMOTO works at the Remote Sensing Technology Center of Japan (RESTEC), Tsukuba, Ibaraki Pref., Japan. She is currently on loan from the Space Engineering Development (SED) Co., Ltd., Tokyo, Japan. She received her B.S. degree in 1993 in Marine Science from Tokai University, Shizuoka Pref., Japan. She is in charge of geometric calibration and characterization of optical sensors on the Advanced Land Observing Satellite (ALOS) including conducting field surveys. Chiyuki NARAMA Chiyuki NARAMA has been a Project Researcher at the Research Institute for Humanity and Nature (RIHN), Kyoto Pref., Japan since 2007, and is working on reconstructing the relationship between environmental changes and human activities in the Balkhash lake basin, Central Eurasia during the last 1,000 years. His profession is physical geography, and he received his Ph.D. in 2002 from Tokyo Metropolitan University, Tokyo, Japan, with a focus on glacier changes in the Tien Shan and Pamir-Alay mountains, Central Asia since the last glacial period. He has been studying glacier changes, glacier hazards, and mountain environmental changes in Central Asia and the Indian Himalayas since Jinro UKITA has been a Professor of the Faculty of Science at Niigata University, Niigata Pref., Japan since He received two M.S. degrees, one in Oceanography from the University of British Columbia and the other Atmospheric Sciences from the University of Washington, U.S., and his Ph.D. from Hokkaido University, Hokkaido Pref., Japan in From 1995 to 1996 he served as a lecturer at the University of Tokyo, and then joined the National Space Development Agency of Japan (NASDA, the former JAXA), where he served as a research scientist and developed a Japan-US joint science program for the International Arctic Research Center (IARC) in Fairbanks Alaska, U.S. as a sub-leader. In 2000 he moved to the U.S. to join the cryosphere programme at NASA Goddard Space Flight Center where he researched sea-ice. Between 1998 and 2004 he was an affiliated research scientist at the Lamont-Doherty Earth Observatory of Columbia University, U.S. Nobuhiro TOMIYAMA Nobuhiro TOMIYAMA has worked at the Remote Sensing Technology Center of Japan (RESTEC), Tokyo Pref., Japan since 2000 on various applications of remote sensing technologies for volcanology, forestry, agriculture, and glaciology. He received his B.E. degree in 1998, and M.E. degree in 2000 from Kumamoto University, Kumamoto Pref., Japan for research into analyses of crustal deformations and topographic changes of active volcanoes using the Interferometric SAR technique. He received his Ph.D. in 2011 as an adult graduate student in the doctoral course at Kumamoto University, Japan. Hironori YABUKI Hironori YABUKI is Team Leader of the Terrestrial Environment Change Research Team, Northern Hemisphere Cryosphere Program, Research Institute for Global Change of the Japan Agency for Marine-Earth Science and Technology (JAMSTEC), Yokosuka, Kanagawa Pref., Japan. He received his M.S. degree from the Graduate School of Science, Nagoya University, and Ph.D. from Nagoya University in He was a researcher of the Frontier Research System for Global Change from 1998 to 2000, a research scientist at the Japan Marine Science and Technology Center from 2000 to 2004, and a research scientist at JAMSTEC from 2004 to (Received 25 October 2011, Accepted 13 February 2012)