1. Introduction

     The purpose of the COSMIC GPS Atmospheric Remote Sensing Field Trip was to provide current U.S students involved and/or interested in GPS atmospheric research the opportunity to learn the latest advances in ground-based and space-based GPS atmospheric remote sensing technology, and to see how such technology is applied in research and operational meteorology in East Asia. In addition, the personal interaction between U.S students and East Asian students and researchers opened doors for increased collaborations between the countries in the future.

     This report describes the scientific activities that took place during the Japan portion of the trip, and the cultural and intellectual experience gained by the U.S. students.

     The first half of the trip consisted of visiting the Japanese cities of Tsukuba, Tokyo, and Kyoto over four days, July 5-8, 2004. Our first stop was Tsukuba, a wonderful, quiet town about an hour outside of Tokyo. We arrived on Sunday afternoon in order to officially start our visit on Monday morning. The group visited the Geographical Survey Institute (GSI) and Meteorological Research Institute (MRI) on Monday. At MRI, some of the U.S students gave short presentations about their current research. Tuesday was filled with tours and talks at the Japan Meteorological Agency (JMA). On Wednesday, the group traveled to Kyoto via the Shinkansen bullet train and visited some area temples. On our last day in Japan, Thursday, we toured the MU Radar facility and a local pottery studio in Shigaraki. Each evening in Japan was an opportunity to further visit with our Japanese colleagues and eat wonderful Japanese cuisine.

2. Geographical Survey Institute

     On the morning of July 5, we visited GSI, the Geographical Survey Institute. GSI is a national surveying and mapping organization involved in all aspects of geographic information, from international surveying activities to formulating policies. As the organization responsible for developing positional and geographical information pertaining to Japan, GSI gathers and processes the data from GEONET sites. GEONET, the GPS Earth Observation Network, is a network of GPS-based control stations covering most of the Japanese Islands. The focus of GEONET is to monitor crustal deformations for geophysical researchers studying earthquakes and volcanic eruptions. In Japan, there are approximately 1224 GEONET stations with an average spacing of 20-25km. On the GSI campus, there were 32 stations. After briefly meeting Dr. Yuki Hatanaka, who works at GSI, he took us to the fields around the building to look at some of the GPS stations.

     Station designs change from year to year, and there was a good sampling of designs on the GSI campus, since not all were installed at once. The exact site chosen for each station was chosen for stability. Each station is composed of a pillar, receiver, antenna, and a modem or terminal adapter. The GPS antenna is mounted on top of the pillar. Inside the pillar are the receiver and modem. Also inside is a grounded battery backup. Some stations had a meteorological package for measuring such things as temperature and pressure.

     Each control station sends its incoming data back to GSI every 30 seconds or 3 hours using IP over telephone lines or satellite communications. At GSI Central Station, the data is converted to RINEX (Receiver-Independent Exchange) format where it is stored in a database and analyzed, displayed, and/or provided to the user community by positioning service providers. After the tour, Dr. Hatanaka gave a short presentation covering more of the details of how GSI processes the data.

3. Meteorological Research Institute

     After visiting GSI, we traveled to the Meteorological Research Institute in Tsukuba. MRI is an associate institute of the Japan Meteorological Agency and was structured in 1977 to include nine research divisions and 27 laboratories.. Current research at MRI focuses on a broad range of areas, from the prediction of volcanic and seismic events to weather forecasting techniques, and scientists at MRI participate in several international research programs like the World Weather Research Program and the Geosphere Biosphere Program.

     While at MRI, we had the opportunity to hear lectures by researchers from seven different institutes and universities, and five of the U. S. students gave presentations about their research interests. Unlike our earlier visit to GSI, where we were able to tour the facilities and see several generations of GPS monuments, the whole of our visit to MRI was devoted to hearing lectures. Still, the interaction that was stimulated by the lectures from Japanese scientists and U. S. students allowed us to better understand some of the current research in Japan; certainly, this interaction complemented very well the knowledge gained from the lectures themselves.

     The first two lectures were given by Dr. Yoshinori Shoji and Dr. Hiromu Seko, both scientists at MRI. Dr. Shoji provided a review of the GPS/MET Japan Project, describing the benefits of GPS remote water vapor sensing to the GEONET and to numerical weather modeling efforts in Japan. Using a four-dimensional variational data assimilation system, precipitable water vapor (PWV) measurements derived from GPS ground station data are assimilated into numerical weather models, in some cases improving the accuracy of forecasts produced by the models. However, statistical results show neutral impact, suggesting that vertical profiling of water vapor, observation of precipitable water in the ocean, and a combination of independent observations from GPS are required to bring out the potential of the dense GPS network. The goal is to use the more accurate forecasts to decrease the position errors in GPS monuments by providing knowledge of the total amount of water vapor along the paths taken by the signals from GPS satellites to the GPS monuments. The amount of water vapor along the signal path affects the time taken by the signal to propagate through the atmosphere, and since measurement of signal arrival time is the basis for determining location using the GPS, accounting more accurately for the signal propagation delay can increase position accuracy. The end result of PWV measurements from GEONET, thus, is an increase in the accuracy of weather forecasts from numerical weather models and a simultaneous improvement in knowledge of the positions of GPS stations - often to within millimeters in the horizontal direction and centimeters in the vertical.

     To further explain some of the topics related to the GPS/MET Japan Project, Dr. Seko provided an evaluation of the use of atmospheric models for GPS data retrieval and a review of water vapor tomography. Elaborating on Dr. Shoji's lecture, Dr. Seko showed how detailed knowledge of atmospheric water vapor content given by numerical weather models allows for more accurate determination of the GPS signal delay, resulting in lower errors in the position of the GPS receiver. Inversely, knowing the signal propagation delay and the precise locations of the GPS satellites and receiver, it is possible to estimate the total water vapor along the signal's path. Dr. Seko described how a grid of GPS receivers on the ground could be used to derive a three-dimensional image of the water vapor in the atmosphere above the grid by breaking the atmosphere into "voxels", or three-dimensional boxes, and applying what is known as the observation equation to each of the boxes.

     Though the GPS-based techniques described by Dr. Seko and Dr. Shoji are not useful for sensing wind velocity, the MU (Middle and Upper) radar, a facility located near Kyoto, can be used as a wind profiler for this purpose. Dr. Toshitaka Tsuda of Kyoto University gave a lecture discussing the MU radar's capability for sensing three-dimensional wind velocities, and also talked about the history and future of space-based GPS radio occultations. An occultation refers to the event when the view of a GPS satellite from a low earth orbiting (LEO) satellite becomes obstructed by the earth, during which time the signal from the GPS satellite passes an increasing distance through the earth's atmosphere. As this happens, the signal's path is bent by the atmosphere, and by measuring the signal delay times from the GPS satellite to the LEO satellite, a refractivity profile of the atmosphere can be derived at the tangent point of the signal path. These refractivity profiles can be translated into profiles of moisture, temperature, and pressure, and can also be used directly to monitor mountain waves and gravity waves.

     The profiles derived from space-based GPS occultations are extremely useful; however, the location and density of these profiles cannot be controlled. Using an aircraft in place of a LEO satellite, it becomes possible to choose the locations of profiles, referred to as down-looking occultation profiles when using this technique, since precise information of all GPS satellites is known in advance. Dr. Takayuki Yoshihara, a researcher at the Electronic Navigation Research Institute, gave a lecture about airborne down-looking occultations and the requisite technology. By placing two GPS antennas on an aircraft, one antenna for GPS positioning and the other for receiving signals from an occulting GPS satellite, it is possible to retrieve profiles at specified locations or times, though it may not always be possible to specify both the time and location of a profile because of the fixed nature of GPS satellite orbits. Dr. Yoshihara explained that aircraft velocities must be known to within about 5 mm/s so that Doppler effects can be distinguished from atmospheric effects on the signal, and to obtain such accurate velocity measurements, the aircraft must be fitted with an inertial navigation system in addition to the GPS receiver for positioning.

     Besides providing profiles of temperature, moisture, and pressure in the troposphere, ground-based GPS techniques can be used to observe properties of the ionosphere. Dr. Takuya Tsugawa of Nagoya University gave a lecture about ionospheric observations using data gathered from GEONET. In his lecture, Dr. Tsugawa explained to us how maps of the total electron content of the ionosphere can be derived from GPS signals received on the ground. Using the dense network of GPS stations that are part of GEONET, it is possible to create maps of total electron content in the ionosphere over Japan every 30 seconds, and with these maps, it is possible to track the wavefronts of disturbances in the ionosphere. Interestingly, we learned that medium-scale traveling ionospheric disturbances, which may be on the order of hundreds of meters in scale and propagate at speeds of about 100 m/s, travel in different directions through the ionosphere, depending on whether it is daytime or nighttime.

     Many of the above applications of the Global Positioning System rely on a very precise knowledge of the positions and velocities of any low earth orbiting satellites that are involved; for example, accurate knowledge of the position and velocity of the LEO satellite is required to gather GPS radio occultation soundings as mentioned above. Dr. Toshimichi Otsubo of the National Institute of Information and Communications Technology described in detail how orbits are calculated with precision beyond what a simple GPS receiver onboard the satellite could provide. By accounting for forces acting on a satellite due to the moon, other planets, ocean tides, drag, and the effects of relativity, it is possible to more accurately compute the acceleration of the satellite and thus determine its orbit more precisely.

     Just as accurate knowledge of LEO satellite orbits is required whenever such satellites are involved, many applications of the GPS involving ground receivers can benefit from accurate knowledge of the position of those receivers as well. Dr. Yuki Kuroishi from the Geographical Survey Institute and Dr. Koji Matsumoto from the National Astronomical Observatory gave lectures about modeling of the geoid for Japan, and continent loading contributions from minor ocean tidal constituents, respectively. In these lectures, Dr. Matsumoto and Dr. Kuroishi explained how accurate models of the geoid, a gravitational equipotential surface of the earth, can be derived from satellite sensors that measure the earth's gravity field, and how factors such as tides and the nutation of the earth can alter the height of points on land.

     After the Japanese scientists described their work, five of the U.S students gave short lectures on their current or planned research. Ron Mastaler, of the University of Arizona, gave a talk entitled "Ground-Based GPS Precipitable Water Observations during the North American Monsoon Experiment (NAME)". NAME was a campaign recently carried out in the southwest by Ron and others working for Dr. Rob Kursinski of the University of Arizona, one of the U.S. leaders in this area of research. Andrew Snyder, of Purdue University, gave a talk entitled "Preliminary Comparison of MODIS and GPS Precipitable Water: Outlook for Coastal Assimilation Studies". His group found that MODIS (Moderate Resolution Imaging Spectroradiometer) provides quality data up until the point where the hurricane clouds interfere with near-IR retrieval technique, and that the temporal resolution of GPS data could be important for hurricanes that intensify at a rapid rate. Dione Rossiter of the University of California-Berkeley gave a talk entitled "Comparison of GPS Radio Occultations and Radiosonde Soundings", a summary of her summer work at UCAR. She found that the radio occultation measurements were accurate enough to detect inconsistencies in radiosonde data. Wallace Hogsett, of the University of Maryland-College Park, gave a talk entitled "Numerical Simulation of Hurricanes". He outlined his future research objectives, which include learning about the energy budgets of hurricanes, analyzing hurricane landfalls using the WRF (Weather Research and Forecast) Model, and investigating the use of GPS data to improve hurricane studies. Jennifer Abernethy, of the University of Colorado-Boulder, gave a talk entitled "A Distributed Heterogeneous Implementation of the WRF Model". She summarized the approach and results of a project to run a forecast with the model domain that is split between two systems across a network; this feature could enable users to get more use out of the different architectures and operating systems they might have in workstations rather than needing one large system that can handle the required domain size and resolution by itself.

4. Japan Meteorological Agency

     The Japan Meteorological Agency (JMA), Japan's version of our National Weather Service, disseminates both short-term and long-term forecast information in Japan. The agency issues short-term weather forecasts, information on earthquake and volcanic activities, and also deals with issues such as global warming.

     JMA meteorologists presented some of their work with both space- and land-based GPS data. In particular, Eiji Ozawa presented results from incorporating the space-based bending angles into their global spectral model using three-dimensional variational data assimilation (3DVAR). The three dimensions of 3DVAR simply include the horizontal (x,y) and vertical (z). The bending angle assimilation was made using CHAMP data, and the goal was to improve forecasts, particularly in the data-sparse Southern Hemisphere. While results here were favorable, they expect that the addition of COSMIC/ROCSAT-3 data will greatly improve model initialization and forecasts over the entire globe.

     Japan's dense network of ground-based GPS monuments, which is made up of both antennas and receivers, has seen widespread use in recent years in the meteorological and geological communities. At the JMA researchers have been assimilating precipitable water vapor data from GSI's GEONET. The high density of the network (~20 km spacing) allows for an excellent sampling of water vapor in both time and space over all of the country's islands. The main difference between 4DVAR and 3DVAR lies in 4DVAR's addition of data over time; therefore, there is a continuous cycling of data into the high-resolution mesoscale model. Any temporal and spatial changes in the moisture content of the atmosphere should be recognized with the assimilation.

     Results from this assimilation have been shown to be inconsistent. At the JMA, the mesoscale model improved its rainfall production with the addition of GPS data, although differences were only slight and nearly insignificant. These results can likely be attributed to the fact that GPS precipitable water is an integration from the ground to the satellite, thus any vertical variation in water vapor will never be resolved. In addition, the source region of moisture over Japan lies in the surrounding bodies of water, where information on vapor content comes from other satellites.

     After a tour of the JMA, described below, Dr. Chris Rocken presented GPS ground based and radio occultation research at NCAR to a group of approximately 60 meteorologists. Dr. Rocken's talk covered both the basics of ground-based and RO retrievals and some of the results of assimilation in the United States. He also iterated that GPS data are attractive because of their independence from other data. This allows for an independent record of the climate, something that will be highly valuable to climate scientists once COSMIC is launched in late 2005.

     During the tour of JMA's forecast lab and Numerical Prediction Division (NPD), several meteorologists in the prediction center showed us model forecasts of recent typhoons and other storms, including Mindulle, a typhoon that devastated portions of Taiwan a week before our visit.

     The tour of the lab also included an overview of JMA, which includes 56 major meteorological stations throughout the islands, with numerous other meso-network stations. The technology used at JMA was rather impressive, as there were numerous large, flat screen monitors, in addition to a vast number of computer workstations.

5. MU Radar

     Thursday, after a traditional Japanese breakfast of sushi and noodles, we traveled outside of Kyoto to visit the MU Radar. The Middle and Upper Atmospheric (MU) Radar instrument is a sophisticated remote sensing tool built in 1984 to collect data on the middle and upper atmosphere. It is operated by the Research Institute for Sustainable Humanosphere of Kyoto University in Kyoto, Japan. The instrument itself is located just outside of Kyoto in Shigaraki, within the lush, hilly landscape.

     Our group was given a tour of the facility by the single full-time on-site scientist, Dr. Mamoru Yamamoto. He explained that, though he is the only full-time scientist who works at the facility, there are frequently many scientists on-site using the instrument by permission of Kyoto University. There are also several women employed at the MU Radar carrying out administrative functions, and they were extremely gracious in using their personal vehicles to drive the members of our delegation up the final winding path to the radar, as our tour bus was unable to negotiate the narrow path and sharp turns.

     In the control and data processing building, we were provided with a simple explanation of how the radar works and how it is controlled by computer from the main building. The MU Radar is unique in that it is an active phased array system. In other words, it is "active" because the instrument sends out pulses of radio energy that bounce off of the atmosphere and return to the radar, providing a picture of its motion. Conversely, a "passive" radar system does not send out a pulse of energy, rather it directly measures the energy given off by a medium. A "phased array" means that the radar is not one single piece of equipment, such as a dish, but a collection of several smaller identical antennae whose measurements are combined.

     We were told that there are many advantages to using such a system to view the middle and upper atmosphere. For example, a radar dish sensitive enough to remotely sense the upper atmosphere would need to be rather large, using 1984 technology when the radar was built. Large dishes can be difficult and/or expensive to operate, as they need to be pointed in multiple directions to function. A phased array, however, can mimic the "pointing" of a dish by adjusting the timing and phase of the pulses emitted by each antenna.

     Following the above explanation, we were lead down to the radar array to see the antennae up close. The array is situated in a shallow basin with a concrete base. There are 475 antennae, and at least an equal number if not more cables connecting the various parts of the array to one of six data collection houses. Inside each house is a computer cluster, which collects data and sends it to the computers in the main processing center.

     The MU Radar is used for a variety of meteorological and astronomical purposes. It captures very high-resolution data of typhoons, frontal passages, three-dimensional winds, and other atmospheric disturbances for use in weather research and forecasting. Additionally, the MU Radar can monitor activity in the ionosphere and provide 24-hour coverage of meteor showers.

6. Comments on the Scientific Experience

     The following are comments made by group members about their scientific experiences in Japan:

     "From the lectures given by the scientists and professors while at MRI, we gained an overview of several different areas of ongoing research in Japan. We learned about current research in GPS applications from scientists working at seven different institutions. Specifically, we learned about space-based GPS research involving GPS satellites and other low earth orbiting satellites, ground-based GPS research using GPS satellites and GPS receivers on the earth's surface, and research in geodesy. Though some of the research described in the lectures may have been hitherto unknown to us, we were able to appreciate the relevance of the research and see its relation to our own research interests on account of our previous familiarity with GPS remote sensing methods."

     "We gained a broader view of how our previous knowledge about GPS remote sensing can be applied or related to research in numerical weather prediction, ionospheric observation, and prediction of the motion of continents."

     "Throughout our time in Japan, we gained a better appreciation for the necessity of much of the science that we heard about in lectures. The amount of volcanic activity and number of earthquakes that Japan is subject to, in addition to the threat of typhoons, creates a demand for science that can enable the prediction of damaging natural phenomena. As an example, we learned that the approximately 1,224 GPS monuments of GEONET enable scientists to track with millimeter accuracy the motion of the different tectonic plates that Japan rests on. Knowledge of the motion of tectonic plates is an asset to earthquake prediction, and further research is partly driven by the need to more accurately forecast seismic events."

7. Cultural Experience and Prospects for Future Collaboration

     This field trip was a whirlwind of scientific visits and learning, but the group did get some time to experience Japanese culture. Increasing our familiarity with the scientific work and culture sparked everyone's interest in establishing collaborations with Japanese scientists in the future.

7.1 Cultural Experiences

     Everyone in the group was very impressed with the Japanese food. Each meal was a cultural experience, because the customs, recipes and presentation were so different than in the U.S. This was everyone's first experience eating sushi for breakfast (in the traditional-style Kyoto hotel) or having a ten-course tofu dinner. Rather than shy away from something 'different', the group looked forward to new experiences and wonderful food at every meal.

     Another cultural experience that everyone enjoyed was the afternoon in Shigaraki. In addition to housing the MU Radar, the town of Shigaraki is the center of a centuries-old pottery industry. The group was given a tour of a traditional pottery production facility and a chance to make our own pieces of pottery that would be fired in their traditional kiln. This provided a unique opportunity for the U.S. delegation to immerse itself in Japanese history. It gave us a glimpse of Japanese history in a more quiet part of the country. We were able to meet people from not only a different culture, but from a different career and lifestyle from our own. As busy budding scientists and/or business people, Shigaraki was a rewarding departure into the world of art and history against the beautiful scenery of the hills.

7.2 Interactions with East Asian Scientists and Prospects for Future Collaborations

     The following are comments from group members about their interactions with the Japanese scientists and prospects for future collaboration:

     "Beyond lectures, at our dinner in Tsukuba, we had the opportunity to talk freely with Japanese researchers about differences in culture, and we were given the chance to share our ideas and opinions on many topics."

     "Although the motivation for some branches of research may be different in Japan, it is apparent that the methods used by Japanese scientists are much the same as in the United States. For this reason, and also considering the wonderful interaction that we had with our hosts and lecturers at all the institutes, we find the idea of collaborating with scientists in East Asia to be very appealing."

     "In any such cooperative work, the contributions from East Asian collaborators would be invaluable; in addition, the resulting cultural interaction would be interesting and enjoyable."

8. Conclusion

     This field trip was a wonderful experience, both personally and professionally, for each member of the group. We realize that we owe a great deal of thanks and respect to our Japanese hosts from GSI, MRI, JMA, MU Radar and Kyoto University, and especially to Bill Kuo, Ted Iwabutchi, Kim Prinzi-Kimbro, Scott Briggs and Chris Rocken of NCAR for all their work in organizing this trip. We greatly appreciate the NSF for funding this opportunity.

     Certainly, all of our visits in Japan and Taiwan have enriched our understanding of science in East Asia, both through visits to institutions and through the informal conversations and dinners that we shared with our hosts and lecturers. Without doubt, our trip has been a wonderful experience that has shaped our views of international collaboration in science and has given us the chance to experience the generosity and cordiality of the people of Japan.