Asian applications of the
'WETNET' concept - A personal computer network to receive daily global
satellite data sets and derived geophysical products James C. Dodge NASA Headquarters, Washington, DC 20546 (USA) Introduction Wet Net is a name given to an experimental NASA computer network that will allow participating scientists to receive daily global satellite images related to studies of physical processes within the Earth's global hydrologic cycle. The remote sensing data will be acquired, processed, and distributed automatically in standardized GIS format images which will allow the scientists to browse the content of numerous satellite channels as well as derived global products such as precipitation rate, atmospheric moisture, cloud liquid water, cloud-free vegetation distribution from the traditional AVHRR "greenness" index and a new microwave "Wetness" index, wind stress over the oceans, land surface and sea surface temperatures, oceanic evaporation, and sea ice extent. Also, accumulated quantities such as estimates of monthly to enable visualization of regional variations. Slowly - varying quantities such as sea ice margins and boundaries of vegetation extent will be illustrated buy making difference maps of mosaic, cloud -free images. If a particular segment of the full-resolution data is desired at the time, the scientist may highlight the region on his computer screen and have it sent to him electronically over phone lines. This mode of high-resolution data transfer is limited to selected regions the global data set because of the data bandwidth of telephone lines. While this system is designed to enable rapid data selection for the study of interacting physical processes within the Earth's global environment, it also; can lead to discovery of unanticipated relationships between the interacting geophysical variables. From an Asian point-of view, a similar system using some of the same methodologies could help in the study of regional environmental interactions and in the solution of problems which extend beyond the geographical boundaries of any particular country. Specifics of the WetNet System In order to evaluate the presumed utility of such a system of daily electronic image delivery to working scientists, we have selected a primary data set and several related ancillary data sets from other satellites observing the Earth during the same time intervals. As the primary data source, we have chosen the microwave imager (SSM/I) carried on all of the DMSP series of U.S. satellites. Its seven channels consist of vertical and horizontal polarizations of 19, 37, and 85 GHz and the vertical polarization of 22 GHz . These particular channels, in various linear combinations, can yield information on various moist processes in the Earth's atmosphere and at the surface. If ancillary information from other satellites, including the U.S. GOES and NOAA satellites as well as the Japanese GMs and Europe's Meteosat, and still other information from surface radar composites is added to the database, then a wide range of algorithms can be developed for various geophysical parameters. With the additional information and available computer power to calculate different algorithms for various scene conditions, latitudes, and seasons, it is presumed that and optimum set of conditional algorithms can be developed for each parameter so that when the set of algorithms is applied to the full global and seasonal data set, the resulting derived physical quantities will have the greatest possible accuracies from remotely sensed data at this time. The reason that such emphasis is being placed on changing physical process involved, and seasonal variations is that we are looking of long-term variations in the Earths environment, so we want to be certain that we have accounted for as many of the remote-sensing algorithm sensitivities as possible. All data from the SSM/I passive microwave sensor will be collected and processed for quality control in order to insure a properly time-ordered and non-redundant data sequence for image generation. The data source at first will be magnetic tapes from NOAA/NESDIS with eventual data flowing, hopefully, over NSFNET in near-real time. The SSM/I data will be received at NASA's Marshall Space Flight Center (MSFC) in Huntsville, Alabama, USA; entered automatically into a large database and assembled for the calculation of numerous daily, weekly, and monthly output product maps. The maps will be transmitted via a scheduler program at the user's terminal so that each scientist need only collect products related to interests. Adequate computer disk space has been provided so that four days worth of data products can reside on the designated disk space available until the oldest data is overwritten. This procedure will allow scientists to return from several days absence from the office and still be able to search the browse products for interesting physical events to study. The transmitted browse products may be selectively saved to permanent memory media such as floppy disks so that either key physical situations are thereby noted for future study or a climatology of browse products themselves is accumulated. In addition to the basic SSM/I data. related ancillary data-- including geostationary images in visible and infrared wavelengths at the times of SSM/I passage over the nadir longitude, manually digitized radar composites over the U.S., and reduced resolution (27 km) AVHRR global data sets -- will be accessed from the satellite data banks of the Space be accessed from the satellite data banks of the Space Science and Engineering Center of the University of Wisconsin and forwarded through a 56 kbaud phone line to the MSFC. Once each day's worth of basic SSM/1 data, related geostationary, and other polar on biter data is assembled , a full set of global and swath data will be used to prepare numerous global browse images in various product formats such as precipitation rate and others listed earlier. Also, collective clouds through compositing for viewing frequently-obscured ground and coastal regions and to evaluate them in terms of vegetation index and sea-surface temperature, respectively. We will also accumulate values of estimated precipitation amount , and average cloud liquid water and precipitately water to look for seasonal behaviour and variability that would help to set a base for consideration of climatic changes from current patterns. There will be an optical disk sent to each scientist every month containing all of the daily data plus all of the derived products and the products prepared from the ancillary data. In this way we can evaluate the strengths and weaknesses of daily data transmission via electronic means compared to monthly dissemination of optical disks. The complete data record in the hands of each researcher also forms what is known as a distributed archive. This methods eliminates the need for future requests of retrospective data. All of the networked scientists will receive the same personal computer system for the receipt and on-site processing of the SSM/I and related ancillary data. It consists of an IBM PS-2, Model 70 with a 120 Mbyte hard disk and 6 Mbytes of RAM. It will be provided with McIDAS image processing software and optical disk reader for archival access as well as a packet-switching modem for achieving wide data bandwidths and error-free data transmissions over conventional telephone lines. The software has been augmented to allow easy menu-driven data handling and mage processing. Even relatively computationally-intensive calculations such as remapping to different map projections can be done directly from menus. Also, strings, of frequently-used commands can be initiated with a single keystroke. Automated receipt of daily data files will be achieved by having an internal software program place the computer in a ready-to-receive mode at a time early during each morning when the telephone lines are used least and when no one is likely to wish to use the PC workstation for research. When the scientist arrives for work each morning, he or she may trigger an automatic or manual scan of the received data. Asian applications of a Wet Net Type of System The advantages and disadvantages of such a system for the Asian community would seem to be as follows. The primary advantages of this technique is that for the selected problem and the agreed remote sensing data sources each participant receives all f the data. So no time or resources are wasted in the process of ordering desired data or data products . Each participants has a sample of all of the data so that he can instantly determine data voids and can use the estimated geophysical parameter maps as a gauge for the quality of the data. While the transmitted maps should not be held to absolute accuracies that would be acceptable in a court of law, they do givc3e an indication how the data channels would respond to the application of a particular estimation algorithm. For the greatest accuracies, a scientist would use the browse products to focus in on a particular data set, then analyze it extremely carefully with his personal computer system to insure that no inadvertent mistakes were made during the calibration, navigation, or channel co-registration stages. He might then also apply several slightly differing algorithms which might account for subtle variations in the particular environmental conditions applicable to the specific scene. Such a system could be used in a single laboratory setting where the analysts are working on the same problems or data sets, but possibility not all working in the same room or building. There, the idea of distributing machine-generated interpretations of routinely-obtained remote sensing data sets to all the persons working on detailed analyses would given the project a degree of standardization for all involved. in the case where the different coworkers are skilled in different scientific disciplines and may not each be experts in each other's remote sensing interpretation techniques, the browse products help to illustrate the varying analysis techniques and serve to provide a cross-training purpose. When one considers such a data transmission system for a team of international scientists, the problems of implementation increase, but not because of a failure of the concept, rather because of the difficulties of reaching international agreements for the standardization of the hardware and software to be used, the analysis algorithm to be used, standard GIS data formats, and methods for effectively and economically using the international communication links. While these difficulties might seem formidable, the advantages of being able to see frequent maps of interpreted data would seem to make the effort worthwhile. If one organization serves as a source of remote sensing imagery for others, this standardized method of providing a data browse image electronically in a GIS format could be an effective way of showing data coverage, quality, and likely interpretations before other data users would order specific high-resolution data. A nominal fee could be charged by the data source organization for the service and the imagery could be provided upon request rather than through routine scheduled distribution. Conclusisssons While a full Asian implementation of an international wetNet type of electronic browse image distribution might be prohibitive due to high recurring communications costs or difficulties in obtaining standardized or compatible hardware in various countries, the value of reviewing satellite image data over the same geographical area on a routine basis as well as related, computer-derived geophysical product maps from that data, would certainly seem to be a convenient way for scientists with different training and located apart from one another to study common environmental problems and to reach consensus conclusions on their causes. |