The Case lor an Expanded Digital WEFAX

The Case lor an Expanded Digital WEFAX Abstract Arguments are presented in support of an intermediate bandwidth rebroadcast service that would make geostationary satellite image data in digital form available frequently to a broader base of users. The appeal of locally available animated cloud motion display is stressed along with the expanding impact of quantitative observables being extracted from digital imagery. By appropriate investment in uplink and rebroadcast facilities, a variety of potential users could acquire data that would be adequate for many operational applications for a fraction of the cost involved in handling present digital image transmissions. 1. Introduction With the response characteristics of earth-sensing vidicon cameras on early ESSA satellites, the narrow-band system for Automatic Picture Transmission provided an adequate imaging capability for relatively simple ground acquisition and display facilities. Once users had invested in such equipment, a natural outgrowth was the compatible Weather Facsimile (WEFAX) retransmission system activated soon after the launch of NASA's first Advanced Technology Satellite (ATS). Since that time, steady and significant progress in developing applications is suggesting the need for a new dissemination approach by which the "essence" of the full digital imaging signals from geostationary satellites can be made available to interface with modest ground facilities. Other technical developments are providing the means that, if implemented, could profoundly enhance the use of satellite data. The present purpose is first to describe the inducements toward such an effort and then to address the means whereby this broader service might be provided. 2. Local need for frequent digital imagery Recent studies have stressed the importance of prompt and frequent satellite data access in the monitoring and short-term prediction of important weather events. Several of these efforts have become experimental operations and other are undergoing further development. One service now supplied routinely by NOAA's National Environmental Satellite Service (NESS) in support of NOAA operations at the World Weather Building is the animated display of SMS/GOES imagery. An automated facility presents a registered sequence of infrared (IR) images with option to select the desired C R T display enhancement. The most recent image is entered and the oldest discarded without operator action, and local 0003-0007 / 78/114-1149$0.00 1978 C. L. Bristor Certified Consulting Meteorologist 309 North Manchester Street Arlington, Va. 223 option permits superposition of an earth locator grid. Remote video displays in some 18 working areas throughout the building support a variety of activities. Insight is provided in support of upper-air analyses, thus permitting more cogent entry of "bogus" observations in sparse data areas. Those involved in quantitative precipitation forecasting receive prompt indications of heavy convective rainfall developments, and severe storm forecasters obtain insight on thunderstorm development prior to meaningful radar returns. Analysts responsible for issuance of snow cover and sea ice charts watch for cloud-free images, and those awaiting pertinent synoptic examples for retrospective studies are aided in their choice. The present facility (SSEC, 1977) displays an 8-picture sequence, but an upgraded system is being considered with 48-frame animation capability. Planning is under way toward the installation of similar animation facilities at National Weather Service (NWS) field sites, but inputs will be by analog land line signals. Although direct interpretation of animated display is mainly qualitative and subjective, user consensus is that one obtains deep insight in support of a wide variety of synoptic applications. With digital input, each user retains flexibility to enhance or otherwise manipulate the displayed information for his particular application. Beyond animation, frequent digital imagery is providing support in several areas of more quantitative applications. In cloud-free regions, surface temperature trends from I R imagery are providing important new inputs for fruit frost forecasting (Southerland and Bartholin 1977). For the monitoring of snow fields, sea ice, and sea surface thermal patterns, transient clouds are eliminated through time compositing (Miller et al, 1977; Waters, 1978), and the integration of cold convective cloud responses from image sequences can provide valuable estimates of convective precipitation (Schofield and Oliver, 1977; Waters et al, 1977a). Extraction of synoptic-scale wind estimates from cloud motion is a well-established operation (Young, 197), and during acquisition of more frequent limited-scan sequences, more detailed mesoscale wind patterns can be obtained (Rodgers et al, 1977). Recent studies using cloud top temperature trends are suggesting I R image sequences as a new tool for monitoring and predicting severe storm development (Adler, 1977; Gentry et al, 1977), and the routine production of a cloud top temperature monitoring chart is now in an operational test phase (Waters et al, 1977b). The demand for frequent digital image transmissions in all of the above applications goes far beyond the 114

1146 Vol. 9, No. 9, September TABLE 1. Likely user candidates for an advanced digital Metrasat service.* User Sites NOAA Weather Service Forecast Offices NOAA Mobile Fire Weather Support Units NOAA Satellite Field Stations (overlaps WSFOs) DOT Air Route Traffic Control Centers DOT U.S. Coast Guard Coastal Centers Other federal agencies State governments College/University meteorology departments Weather-related research institutes Operational weather service companies Industries with internal weather support units Fisheries (vessel or shore sites) Coastal shipping and offshore drilling T V stations with major coverage Cable TV companies Commodity marketing companies Domestic and foreign airlines Western Hemisphere countries Experimenters, radio/computer amateurs, etc. Potential Possible 19 1 0 3 2 6 3 1 2 7 0 0 2 30 * The number of potential users is based on the following sources: Bates (1976), Bureau of Census (1977), Business Week (1978), Krich and Sussman (1976), NWS (1977), Ramage (1978), and Shuman (1978). In a few instances, it was based on personal contacts. Magnitudes are regarded as conservative. The number of possible users is more speculative. S-band W E F A X service now developing. Even for other quantitative applications that might use redigitized W E F A X imagery, the quality of the limited bandwidth analog signal and the availability of coverage for a specific area at only 3 h intervals creates substantial roadblocks for the user. At present, the only recourse involves installation of more costly acquisition facilities to acquire the full "stretched" V I S S R (visible and infrared spin scan radiometer) signals. Although compromise solutions are possible (Bristor and Raynore, 1977), strong appeal could be made to a much broader user community if initial investments could be limited to a modest augmentation of an S-band W E F A X acquisition facility. W i t h recent advances in space communications technology and the availability of inexpensive data acquisition and processing equipment, it is urged that we reappraise the present program and explore such an opportunity. Projecting an advanced quantitative data applications payoff combined with modest costs, one can envision a profound broadening of the user base. According to past surveys, there are 6 university and college departments of meteorology in some 37 states, and private industrial meteorologists provide operational forecast services in some 2 centers in 18 states. C o m b i n i n g these with a variety of other potential user groups, one obtains a rather impressive list ( T a b l e 1). If only those listed as 1978 "possible" constituted the initial user group, a good beginning of the broadening trend would have been made. 3. An intermediate system T h e suggested system represents a compromise between the present W E F A X and the much higher bandwidth stretched V I S S R rebroadcasts. T h e essence of the digital imagery would be retransmitted at intermediate bandwidth by eliminating space-viewed response, by cropping foreshortened edge portions, by removing overlapped samples, and by restricting the transmission of highestresolution visible channel data to only those areas of prime synoptic significance. Using the present stretched V I S S R signal as input, it is proposed that additional ground facilities and perhaps leased satellite bandwidth provide this Metrasat ( Transmission Satellite) service. A critical factor involves the use of greater spacecraft transponder resources so that a digital signal at intermediate bit rate could be received with essentially the same ground facilities now employed for conventional W E F A X transmissions. W i t h rebroadcast information delimited to an optimized subset of the stretched V I S S R content, such an intermediate service could provide most of the information required by the majority of users. Employing a 60 X * bps (bits per second) data stream, for example, one might project 14 min staggered rebroadcasts from the two operational S M S / G O E S imaging satellites. At this rate, 36 X 3 bits could be transmitted during each 600 ms spin period, and one might deploy the available bit volume as follows: Type of Information Bits I R samples (8 km, nonoverlap, with 9th grid bit) 20 Visible channel samples (4 km, 6 bit) Visible channel samples (1 km, 6 bit) Synchronizing and documentation bits Total: 130 130 9000 300 36000 In 14 min, all such imagery could be made available for the areas indicated in Fig. 1. T h e start of transmission would be delayed slightly until data for the southernmost sector were acquired. Many other combinations are possible, and the exact mix would be designed to optimize service to the user community. As under present stretched V I S S R operations, when limited-scan imagery is acquired, service would be restricted to information under that acquisition mode. At night, when only I R imagery is acquired, other information could be rebroadcast such as A V H R R (advanced very high resolution radiometer) imagery from T I R O S - N, derived graphic or alphanumeric satellite products (winds, soundings, etc.), or N W S products (analyses, prognostic charts, etc.). Such service, with significant bit rate reduction (as compared with present stretched V I S S R rebroadcasts), retains the salient image feature information of m a x i m u m utility for most quantitative applications.

1147 FIG. 1. Imagery obtainable in 14 min under the proposed Metrasat system. A favorable power-bandwidth relationship is needed to receive the digital signal with a modest antenna. U n d e r the proposed system, this would be provided by using a power level similar to that employed in present stretched V I S S R rebroadcasts but with the bandwidth reduced from that required for the present 1.7 X 6 bps to the projected 60 X s bps rate. T h e resulting advantage in ground signal-to-noise ratio would permit use of a 1.2 m parabolic antenna with low-cost preamplifier and downconverter components. A simple, fixed antenna mount would permit signal acquisition even with a meandering satellite (imperfect orbital inclination) because of the relatively wide beam width. All components between the antenna and receiver can be assembled and integrated with little electronic expertise and can be made available for approximately $00 (Microcomm, 1976). Based upon preliminary checks with several likely sources, the total acquisition system (including receiver, power supplies, signal demodulator, b i t / f r a m e synchronizer, and antenna with the above-mentioned frontend components) can be purchased for about $6000. T h e basic digital data handling components for a beginning system would include a microprocessor with incoming bit stream access channel, flexible disk storage, keyboard/printer, and a simple C R T image display. Because a local user would likely have interest in only limited subsets of each image transmission, this would provide adequate data buffering storage and a considerable manipulative capability. Output would be restricted to static image display or printed image arrays and other graphic or alphanumeric derivations. Even so, such a facility can support many meaningful quantitative applications for a cost of approximately $8000. Interactive animated display systems are presently considerably more costly, but minimum-priced units using a standard color T V for display appear feasible.

1148 With an initial engineering design investment of about $ 000, it appears that presently available refresh memory systems might be modified to provide 4-frame animation for a unit price under $4000. With the approaching availability of higher-density solid state memory chips, an 8-frame animation system could likely be offered for less than $8000. Such a minimum cost C R T and video animation system could provide more advanced trend and cloud motion measurement capability. For users without local technical resources, the system integration, installation, and software assistance could cost an additional $00-$ 000. Those having in-house talent and some computer support already available, could likely reduce costs for a system with animation and cloud motion extraction capability to a total expenditure of $2 000-$30 000. Even though such a minimum cost configuration may be regarded as an experimental facility, it nonetheless can provide a meaningful performance capability. Users with larger budgets and greater benefitto-cost considerations would have adequate source data to justify expanded facilities with greater capabilities, and network-oriented users could economically provide units at individual sites with "stand-alone" capability and so alleviate land line costs and analog signal dissemination problems. 4. Getting started Despite speculation as to the extent of beginning user involvement, the proposed approach has great potential for broadening the base of quantitative users of satellite image data. Many universities and colleges could devise a beginning acquisition and processing system as a core feature in a synoptic meteorology curriculum, and the acquired data sets, after timely laboratory use, would provide inputs for a variety of study projects. Operational agencies with "nowcasting" or very-short-range forecasting missions would be able to assess the current status of critical weather events and make short-term projections. Others with crop survey or other missions requiring prolonged data accumulations would be able to obtain direct and timely, high-quality inputs for their data bases. All of these goals could be achieved at a small fraction of the cost of user facilities now required to receive and manipulate the full VISSR signals. As a feasibility and validation test, it is proposed that such Metrasat experiments be carried out using the transponder capabilities on a spare SMS/GOES spacecraft. The full capability for stretched VISSR transmission would be used in rebroadcasting the 60 X s bps signal. A test acquisition facility with a small antenna and with appropriate band-pass capability would validate ground system costs and confirm receipt of test signals with acceptable bit error rate. For the uplink signal, an experimental minicomputer with suitable modulator and transmitter interface would be required. Alternately, an existing NOAA Synchronizer/Data Buffer with experimental software might be diverted for brief periods (say, as a substitute for an occasional Vol. 9, No. 9, September 1978 single-frame WEFAX transmission) in support of such limited tests. Once the feasibility of the intermediate system is established, full activation might be achieved using a spare GOES spacecraft or a commercial communications satellite facility. Using present stretched VISSR signals as input will, of course, require separate buffering and uplink facilities for such an operation. In view of the potential for such a readily available digital image source for interested government agencies alone, the cost of the rebroadcast signal could likely be equated to the reduced cost of needed additional field acquisition systems. The broadened base of nongovernmental users would provide expanded opportunities for data use and thus enhance the entire program in terms of taxpayer investment. Private enterprise efforts would be encouraged, and a more effective inter- rebroadcast system would have been established. Bringing an acquisition system within reach of radio amateurs and experimenters would also encourage even less costly ground system developments (Shuch, 1977; Taggart, 1977), and the growing group of serious computer hobbyists might further assist in information usage developments.. Alternatives and recommendations Considering the likely million dollar cost of leased bandwidth for a commercial satellite data relay service, one should, of course, explore alternatives. Amending present WEFAX service by shifting to a digital transmission mode might first be considered. But with present bandwidth constraints, the realizable bit rate is completely inadequate. Increasing the bandwidth of a concurrent WEFAX/VISSR data relay/acquisition operation would likely create other problems. The resulting power budget strain, the likely worsening of cross-talk conditions presently encountered, and the growing inducement to maintain international uniformity between present WEFAX transmissions would all seem to preclude this as a viable solution. One might otherwise suggest amending the present stretched VISSR retransmissions by cropping space response and otherwise eliminating bulk so as to reduce bit rate. But the required reduction in visible channel spatial resolution would seriously hamper those already committed to full acquisition and more complete use of the full data bulk. Another possibility involves using a spare GOES spacecraft in a modified Mode C retransmission configuration (NESS, 1972). However, since the 33 X 3bps data rate in that mode involves only uplink signal considerations, adherence to such a rate presents no obvious advantage. Clearly, the basic proposal should be optimized in favor of low-cost ground acquisition facilities as a function of spacecraft power/bandwidth considerations and in terms of signal modulation alternatives (Lucky et al., 1968). Several amendments to the proposal can also be suggested in order to attempt further benefit/cost advantages for the proposed system. One alteration would

minimize overlap between the dual east-west satellite imaging coverage. T h e tradeoff could then provide greater latitudinal coverage or larger 1 km resolution sectors in the visible imaging channel. Another possibility involves data compression. Predictor algorithms that compress bit volume by using only immediately preceding samples can supply twice the sample population in the same bit bulk with perfect recovery. Other compression algorithms require an image stripe consisting of three or five image lines to permit cluster compression treatment. Such techniques can reduce bulk at least by a factor of 4. Although such capability could be incorporated in the uplink data handling facility at modest cost, user reaction would likely be mixed. Those wishing to copy only limited subsets of the data strings could not use simple on-line cropping procedures because of the encoding complications but would be required to ingest the entire bit stream during each scan period. Unless a majority of users agreed to address the reexpansion task and endure the additional buffering penalty, this possibility would not be acceptable. T h e more costly but simpler alternative would provide more data by continuing the transmission from each V I S S R source for the full 30 min acquisition cycle. Elimination of the 14 min alternating time slots would, of course, require a doubling of the communication satellite resources. Other alternatives are no doubt possible, and one might also consider different combinations of the above suggestions. Before inauguration of any such intermediate service, users should be polled for an expression of preferences and overall interest. Ideally, NOAA's Coordinator for Direct Readout Services might conduct a survey by questionnaire in order to ensure an optimum approach. Past surveys (e.g., Bureau of the Census, 1977) indicate a reluctance on the part of nongovernmental users to make substantial investments for weather data acquisition. However, the above proposal would provide the user with the means for monitoring and projecting weather events on a continuous basis and on a spatial scale not achievable in any other way. Acknowledgments. Valuable suggetsions were provided by B. Clemson, P. Shuch, P. Taggart, and J. Puerner. References Adler, R., 1977: Satellite-based thunderstorm intensity parameters. Preprints, th Conference on Severe Local Storms (Omaha), AMS, Boston, pp. 8-1. Bates, C. C., 1976: Industrial meteorology and the. Bull. Amer. Meteor. Soc., 7, 131327. Bristor, C. L., and W. L. Raynore, 1977: Digital satellite imagery in industrial meteorology. Bull. Amer. Meteor. Soc., 8, 480-487. Bureau of the Census, 1977: Weather data needs survey. Conducted for NWS, NOAA, Washington, D.C. Business Week, 1978: Perverse weather The boom in private forecasting. Feb. 27, 60-64. 1149 Gentry, R. C., E. Rodgers, J. Steranka, and W. E. Shenk, 1977: Equivalent blackbody temperatures of cloud tops and tropical cyclone intensity. Preprints, 11th Technical Conference on Hurricanes and Tropical Meteorology (Miami), AMS, Boston, pp. 274-279. Krich, S. I., and S. M. Sussman, 1976: A concept and plan for the development of a weather subsystem for air traffic control. Project report, FAA-RD-76-23, for Federal Aviation Administration, Lincoln Lab., Lexington, Mass. Lucky, R. W., J. Salz, and E. J. Weldon, Jr., 1968: Principles of Data Communications. McGraw-Hill, New York, pp. 246-276. Microcomm, 1976: S-band WEFAX receiver downconverter circuit modules. Price list and specification. (Available from Microcomm, 14908 Sandy Lane, San Jose, Calif. 9124.) Miller, D. B., et al., 1977: Potential applications of digital visible and infrared data from geostationary environmental satellites. Paper presented at the 11th Remote Sensing Symposium, Environmental Research Institute of Michigan, Ann Arbor. NESS, 1972: Receiving system, data utilization station. Specification No. S24.023, National Environmental Satellite Service, NOAA, Washington, D.C. NWS, 1977: Operations of the National Weather Service. NOAA/NWS Publ., Silver Spring, Md. Ramage, C. S., 1978: Further outlook Hazy. Bull. Amer. Meteor. Soc., 9, 18-21. Rodgers, E., R. C. Gentry, W. E. Shenk, and V. J. Oliver, 1977: The benefits of using short interval satellite imagery to derive winds for tropical cyclones. Preprints, 11th Technical Conference on Hurricanes and Tropical Meteorology (Miami), AMS, Boston, pp. 63-660. Scofield, R. A., and V. J. Oliver, 1977: A scheme for estimating convective rainfall from satellite imagery. NOAA Tech. Memo. NESS 86, Washington, D.C. Shuch, H. P., 1977: A cost-effective modular downconverter for S-band WEFAX reception. IEEE Trans. Microwave Theory Tech., T - M T T - 2, 1127-1131. Shuman, F. G., 1978: Numerical weather prediction. Bull. Amer. Meteor. Soc., 9, -17. Sutherland, R. A., and J. F. Bartholic, 1977: A freeze forecasting model based upon meteorological satellite data. Preprints, 13th Agriculture and Forest Meteorlogy Conference (West Lafayette, Ind.), AMS, Boston, pp. 4-46. SSEC, 1977: Satellite image sequencer operating instructions. Report, DOC contract 3-3372, for NESS/NOAA, Space Science and Engineering Center, Univ. of Wisconsin, Madison. Taggart, R. E., 1977: An S-band receiving station for the GOES weather satellites. 73 (Radio Amateur) Magazine. Waters, M. P., Ill, 1978: Operational and experimental use of SMS/GOES digital satellite data. Paper presented at the 12th Remote Sensing Symposium, Manila, Environmental Research Institute of Michigan, Ann Arbor., C. G. Griffith, and W. L. Woodley, 1977a: Use of digital geostationary satellite imagery for real-time estimation of hurricane rain potential in landfalling storms. Preprints, 11th Technical Conference on Hurricanes and Tropical Meteorology (Miami), AMS, Boston, pp. 198-3., D. B. Miller, D. C. Dismachek, and R. M. Carey, 1977b: Cloud height and change analysis from digital SMS/GOES satellite data. Preprints, th Conference on Severe Local Storms (Omaha), AMS, Boston, pp. 42-47. Young, M. T., 197: The GOES wind operation. NOAA Tech. Memo. NESS 64, Washington, D.C.