g. M. isr: e*l ni-us >o^ 
 
 ESSA TR ERL 171 -ITS 109 
 
 A UNITED STATES 
 DEPARTMENT OF 
 COMMERCE 
 PUBLICATION 
 
 <** °'C<v 
 
 V^*-'.*« r 
 
 ESSA Technical Report ERL 171 ITS 109 
 
 U.S. DEPARTMENT OF COMMERCE 
 Environmental Science Services Administration 
 Research Laboratories 
 
 The Basis and Application 
 
 of a Simulation Model for an Oceanographic 
 
 Data Transmission System Using HF Radio 
 
 DALE N. HATFIELD 
 JEAN E. ADAMS 
 
 BOULDER, COLO. 
 JUNE 1970 
 
ESSA RESEARCH LABORATORIES 
 
 The mission of the Research Laboratories is to study the oceans, inland waters, the 
 lower and upper atmosphere, the space environment, and the earth, in search of the under- 
 standing needed to provide more useful services in improving man's prospects for survival 
 as influenced by the physical environment. Laboratories contributing to these studies are: 
 
 Earth Sciences Laboratories: Geomagnetism, seismology, geodesy, and related earth 
 sciences; earthquake processes, internal structure and accurate figure of the Earth, and 
 distribution of the Earth's mass. 
 
 Atlantic Oceanographic and Meteorological Laboratories: Oceanography, with emphasis 
 on the geology and geophysics of ocean basins, oceanic processes, sea-air interactions, 
 hurricane research, and weather modification (Miami, Florida). 
 
 Pacific Oceanographic Laboratories: Oceanography; geology and geophysics of the 
 Pacific Basin and margins; oceanic processes and dynamics; tsunami generation, propaga- 
 tion, modification, detection, and monitoring (Seattle, Washington). 
 
 Atmospheric Physics and Chemistry Laboratory: Cloud physics and precipitation; chem- 
 ical composition and nucleating substances in the lower atmosphere; and laboratory and 
 field experiments toward developing feasible methods of weather modification. 
 
 Air Resources Laboratories: Diffusion, transport, and dissipation of atmospheric con- 
 taminants; development of methods for prediction and control of atmospheric pollution 
 (Silver Spring, Maryland). 
 
 Geophysical Fluid Dynamics Laboratory: Dynamics and physics of geophysical fluid 
 systems; development of a theoretical basis, through mathematical modeling and computer 
 simulation, for the behavior and properties of the atmosphere and the oceans (Princeton, 
 New Jersey). 
 
 National Severe Storms Laboratory: Tornadoes, squall lines, thunderstorms, and other 
 severe local convective phenomena toward achieving improved methods of forecasting, 
 detecting, and providing advance warnings (Norman, Oklahoma). 
 
 Space Disturbances Laboratory: Nature, behavior, and mechanisms of space disturb- 
 ances; development and use of techniques for continuous monitoring and early detection 
 and reporting of important disturbances. 
 
 Aeronomy Laboratory: Theoretical, laboratory, rocket, and satellite studies of the 
 physical and chemical processes controlling the ionosphere and exosphere of the earth and 
 other planets. 
 
 Wave Propagation Laboratory: Development of new methods for remote sensing of the 
 geophysical environment; special emphasis on propagation of sound waves, and electro- 
 magnetic waves at millimeter, infrared, and optical frequencies. 
 
 Institute for Telecommunication Sciences: Central federal agency for research and 
 services in propagation of radio waves, radio properties of the earth and its atmosphere, 
 nature of radio noise and interference, information transmission and antennas, and meth- 
 ods for the more effective use of the radio spectrum for telecommunications. 
 
 Research Flight Facility: Outfits and operates aircraft specially instrumented for re- 
 search; and meets needs of ESSA and other groups for environmental measurements for 
 aircraft (Miami, Florida). 
 
 ENVIRONMENTAL SCIENCE SERVICES ADMINISTRATION 
 
 BOULDER, COLORADO 80302 
 
.-(■^eNTo^ 
 
 Sc 'enct »«*#* 
 
 U.S. DEPARTMENT OF COMMERCE 
 Maurice H. Stans, Secretary 
 
 ENVIRONMENTAL SCIENCE SERVICES ADMINISTRATION 
 
 Robert M. White, Administrator 
 RESEARCH LABORATORIES 
 Wilmot N. Hess, Director 
 
 ESSA TECHNICAL REPORT ERL 171-ITS 109 
 
 The Basis and Application 
 
 of a Simulation Model for an Oceanographic 
 
 Data Transmission System Using HF Radio 
 
 DALE N. HATFIELD 
 JEAN E. ADAMS 
 
 a 
 o 
 
 c 
 
 a 
 
 v 
 o 
 
 INSTITUTE FOR TELECOMMUNICATION SCIENCES 
 BOULDER, COLORADO 
 June 1970 
 
 For sale by the Superintendent of Documents, U. S. Government Printing Office, Washington, D. C. 20402 
 
 Price 35 cents 
 
FOREWORD 
 
 The work described in this report was done at the request of the 
 Commandant, U.S. Coast Guard. It was funded under Interdepart- 
 mental Purchase Request No. Z-70099-9-91574, dated 16 September 
 1968, and monitored by LCDR Marshall E. Gilbert, USCG. 
 
TABLE OF CONTENTS 
 
 Page 
 
 ABSTRACT 1 
 
 1. INTRODUCTION 1 
 
 1. 1 Background 1 
 
 1. 2 Characteristics and Prediction of HF Circuit 
 
 Performance 3 
 
 2. IGOSS BASELINE TELEMETRY SYSTEM 6 
 
 3. TELEMETRY SYSTEM PERFORMANCE 10 
 
 4. PROGRAM DESCRIPTION 12 
 
 5. SAMPLE RESULTS 16 
 
 6. CONCLUSIONS 23 
 
 7. ACKNOWLEDGMENTS 23 
 
 8. REFERENCES 24 
 APPENDIX A. Description of Subroutines Used by PREDPRFM 25 
 
 in 
 
THE BASIS AND APPLICATION OF A SIMULATION MODEL 
 FOR AN OCEANOGRAPHIC DATA TRANSMISSION SYSTEM 
 
 USING HF RADIO 
 
 Dale N. Hatfield and Jean E. Adams 
 
 This report describes a digital computer program for 
 simulating ocean data transmission systems that use 
 high frequency radio. It is intended for use in the design 
 and operation of the Integrated Global Ocean Station 
 System (IGOSS) and the National Data Buoy Systems. 
 Requirements for these two systems are reviewed, and 
 important factors affecting overall transmission system 
 performance are identified. The program predicts 
 performance as a function of these parameters for use 
 in cost/effectiveness studies. A baseline system is 
 described, and sample results from the program are 
 given for it. 
 
 Key words: Digital computer simulation, HF telemetry 
 system, Integrated Global Ocean Station System, 
 National Data Buoy Systems, oceanographic data trans- 
 mission, performance predictions. 
 
 1. INTRODUCTION 
 
 This report describes a digital computer based simulation model for a 
 high frequency (HF) data transmission system. Section 1. 1 gives back- 
 ground information on the development of the Integrated Global Ocean 
 Station System and the National Data Buoy Systems, which the data 
 transmission system will serve. A brief summary of the characteristics 
 and the methods used for predicting the performance of HF communi- 
 cation systems is given in section 1. 2. 
 
 1. 1 Background 
 
 After World War II, it was recognized that a large amount of marine 
 environmental data were needed for predictions of weather, fish 
 movements, etc. At the same time, automated data handling systems 
 were being perfected, which enabled processing of large amounts of 
 
data, thus providing a means for processing the data from a large scale 
 ocean data system. 
 
 An interrelationship between the upper ocean and the lower atmosphere 
 has been long recognized. Consequently, the atmosphere and the ocean 
 should be considered as interacting parts of a single unified global fluid 
 system. A comprehensive understanding of the interplay is essential in 
 describing and predicting the states of the ocean-atmosphere environment. 
 The need to understand and predict this environment necessitates the 
 timely collection and dissemination of these data on a greatly expanded 
 global basis. Considering that about 70 percent of the earth's surface is 
 covered by oceans, and that these have traditionally been a sparse data 
 zone, one sees the importance of a global ocean data system. 
 
 Such a system will contribute not only to the enhancement and safety of 
 men and material, but will also be of great economic value to the fishing 
 industry, ocean floor exploitation, and improved overall weather fore- 
 casting. It must be international in scope because a purely national 
 development would probably be inadequate as well as inefficient. Recog- 
 nition of this led to the development of the Integrated Global Ocean Station 
 System (IGOSS) concept by the Intergovernmental Oceanographic Committee 
 (IOC) that calls for just such international cooperation in gathering and 
 exchanging of marine meteorological and oceanographic data. 
 
 At the national level, a similar concept of cooperation among Federal 
 agencies, academic groups, and private research establishments 
 evolved. The U. S. Coast Guard was given the responsibility for 
 developing the national oceanographic data collection system. The 
 National Data Buoy Development Project at U. S. Coast Guard Head- 
 quarters is the center for this development. 
 
 In 1962, the IOC recommended that a joint World Meteorological 
 Organization (WMO) and IOC working group be established to explore 
 in detail the gathering of data from platforms on the oceans. 
 
 The WMO, of course, has long been engaged in efforts to unify the 
 national meteorological services on a global basis, notably through 
 planning the World Weather Watch system. From joint WMO/IOC 
 efforts, the World Administrative Radio Conference (held in Geneva in 
 1967) allocated six 3. 5 kHz bands in the HF radio spectrum for ocean 
 data transmission. The United States supported these proposals, and 
 the Institute for Telecommunication Sciences (ITS) prepared a detailed 
 analysis (Haydon, 1967) demonstrating the feasibility of HF radio trans- 
 mission for a global system. 
 
In 1968, ITS was given the responsibility to make further studies of the 
 data transmission system using the six allocated frequency bands. The 
 objectives were: 
 
 (1) to review and make recommendations of system parameters, 
 
 e. g. , modulation techniques, data rate, antennas, and required 
 transmitting power. 
 
 (2) to develop a plan for the coordinated use of the allocated 
 spectrum. 
 
 These objectives were to be considered from an international and 
 national standpoint. 
 
 Since the National Data Buoy System is an autonomous subset of the 
 international system, the questions posed at the national level are 
 paralleled by questions at the international level. The ultimate size of 
 the system and the complexity and variability of HF propagation led to 
 developing a digital computer model of the communication system. This 
 report describes a preliminary version of this model, which should be 
 useful not only for designing the communication system, but also for 
 developing and evaluating national /international coordination procedures. 
 
 1. 2 Characteristics and Prediction of High 
 Frequency Circuit Performance 
 
 The oceanographic data transmission system can be considered as a 
 large number of individual radio circuits that must be coordinated (a 
 circuit is defined as the path from one platform to one shore station). 
 The performance of the total system is obviously dependent on the per- 
 formance of these individual circuits. These circuits will make use of 
 the six frequency bands that were allocated by the ITU. These frequency 
 bands are: 
 
 4, 162. 5-4, 166. kHz 
 
 6,244. 5 - 6,248. 
 
 8, 328. 0-8, 331. 5 
 12,479. 5 - 12,483. 
 16,636. 5 - 16,640. 
 22, 160. 5 - 22, 164. 
 
 The performance (say in terms of error rate for a digital transmission 
 system) on one of these circuits on one of the allocated bands is a 
 highly variable factor. Fortunately this performance is predictable, in 
 
a probabalistic sense, over a period of a month. That is, the percentage 
 of days of the month that a required performance level will be met or 
 exceeded at a specified hour can be calculated. This percentage of days 
 is frequently referred to as the circuit reliability for that hour during 
 the month, and it can be thought of as being the product of two 
 probabilities: 
 
 (1) the probability that the ionosphere will support (reflect) the 
 signal at that hour, and 
 
 (2) if it is reflected, the probability that it will be strong enough 
 to give the desired performance. 
 
 Both probabilities depend upon the intensity and distribution of ionization 
 of the ionosphere along the path. 
 
 The first factor, the probability of support, is essentially independent 
 of circuit design parameters (e. g. , transmitter power level). It 
 depends basically on the operating frequency, the angle of incidence on 
 the reflecting layer, and the intensity of ionization. The higher the 
 frequency and/or the steeper the angle of incidence, the more intense 
 the ionization must be to reflect the signal. Thus for a given path 
 geometry, i. e. , angle of incidence, the probability of support decreases 
 with increasing, frequency. For example, there maybe only a 10% 
 chance of a 22 MHz signal being supported, but a 95% chance for a 4 MHz 
 signal. Also because of the steeper angles of incidence associated with 
 shorter paths, the probability of support tends to decrease with decreasing 
 path length. The ionization of the ionosphere is produced by solar radiation 
 and the intensity of ionization depends on local time of day along the path, 
 season, and general level of solar activity. The intensity, and hence 
 probability of support, increases during daylight hours and during years 
 of high solar activity. As a result, higher frequencies are more useful 
 during these periods. 
 
 The second factor, the probability of adequate signal strength, depends 
 upon (1) the ionization in the lower regions of the ionosphere along the 
 path, (2) the geometry of the path, (3) equipment parameters and per- 
 formance quality desired, (4) operating frequency, and (5) the radio 
 noise and interference environment at the receiver site. HF signals 
 are absorbed in passing through the lower regions of the ionosphere on 
 their way to or from the higher regions where reflection normally takes 
 place. The amount of this absorption increases with distance travelled 
 through the lower regions and with the intensity of ionization present in 
 them. Thus the signal strength and the probability of an adequate 
 
signal decreases on longer paths, during daylight hours, during the 
 summertime, and during years of higher solar activity. The absorption 
 is inversely proportional to frequency, that is, lowering the operating 
 frequency increases the absorption. The probability of adequate signal 
 strength then decreases with decreasing frequency, while the first 
 factor, the probability of support, increases with decreasing frequency. 
 
 In general, the curve of reliability versus operating frequency for a 
 given time and set of parameters tends to have a peak because of these 
 opposing propagation related factors. The circuit performance on 
 frequencies below this peak is degraded faster due to absorption than it 
 is improved due to a higher probability of propagation. Conversely, for 
 frequencies higher than this peak, the probability of support drops faster 
 than the probability of adequate signal increases. 
 
 The adequacy of the signal at the receiver also depends on the radio noise 
 level on the operating frequency. This noise may be man-made (such as 
 ignition noise from automobiles), atmospheric (lightning discharges) 
 galactic, or, in rare instances in these bands, noise in the receiver 
 itself. As an example, high man-made radio noise levels at urban 
 receiving areas would require greater signal strengths than those at 
 rural receiving areas for the same level of performance. Signal 
 strengths increase with transmitter power and antenna gain. Thus 
 higher radio noise levels tend to narrow the band of useful frequencies 
 while higher transmitter powers and antenna gains tend to increase it 
 by increasing the probability of an adequate signal. 
 
 A final consideration in factor two is the required signal-to-noise ratio 
 (SNR) for the desired level of performance with the modem and coding 
 used. In a digital data collection system, the major measure of per- 
 formance is specified by the bit error rate. Higher allowable error 
 rates require lower SNR, and expands the useful frequency band. 
 
 To summarize, the frequency at which the reliability is maximum and 
 the value at that point of maximum reliability, as well as overall shape 
 of the reliability versus frequency curve will change with geographic 
 location, time of day, season, solar activity, equipment parameters, 
 and required circuit performance. For best performance the operation 
 of an individual circuit should be on a frequency at or near this point of 
 maximum reliability. 
 
 Calculating HF circuit performance is a long, tedious process if manual 
 methods are used. Recently computer programs have been developed 
 
that can compute this reliability rapidly. The computer program 
 originally developed at ITS is described in Lucas and Haydon (1966), 
 and the latest improved version called HFMUFES, which has been 
 adapted for this simulation program, is described in Barghausen et al. 
 (1969). The program uses predictions of the sun's activity and basic 
 ionospheric parameters that are obtained from long term collection of 
 vertical incidence soundings, and atmospheric and man-made radio 
 noise measurements. Both the ionospheric and noise data are supplied 
 to the program in the form of numerical coefficients defining numerical 
 map functions (Jones et al. , 1966). From these functions the parameter 
 of interest can be determined as a function of geographic coordinates 
 and time. The computer program uses the predicted ionospheric 
 parameters to predict propagation modes, probability of support, 
 signal strengths, and other circuit factors. The atmospheric noise 
 maps, and assumptions regarding the man-made noise environment, 
 are used in the program to calculate the noise power at a given frequency, 
 time, and receiver location. Then, from the predicted median signal 
 power, the median noise power, the variances, and the required signal- 
 to-noise ratio, the probability of an adequate signal is computed. The 
 reliability is then computed. 
 
 2. IGOSS BASELINE TELEMETRY SYSTEM 
 
 This section describes the basic IGOSS communication system around 
 which the simulation computer model was developed. This baseline 
 system serves as a point of departure for the communication system 
 analysis. The communication system is to efficiently transfer oceano- 
 graphic and meteorological data collected by the ocean station to the 
 shore where it can be processed and passed on to the users. 
 
 The IGOSS concept projects an ultimate system consisting of several 
 types of ocean stations (platforms), including moored and unmoored 
 buoys, instrumented merchant ships, and research vessels. For the 
 baseline system, a group of several hundred buoys and a thousand ships 
 might be considered. Platforms could be located in any ocean of the 
 world, so no specific geographic distribution is inherent in the assumptions 
 for the baseline system. The communication system performance will 
 depend on the distribution of platforms. The importance of this point 
 will be discussed later. 
 
 Data from the platforms are sent to shore stations that are appropriately 
 equipped. The number and location of shore stations will affect the 
 communication system performance and cost. For the initial system a 
 group of 18 shore stations was selected such that any ocean area is 
 
within 4, 000 km of at least one shore station. The 4, 000 km is about 
 the maximum distance for a one hop propagation path. The ocean area 
 within this 4, 000 km radius will be referred to as the primary coverage 
 area. Wherever possible, the shore station sites were chosen at 
 existing World Weather Watch facilities. The distribution of these 18 
 stations is shown in figure 1. It must be emphasized that this is an 
 initial assumption for a baseline system only, and the actual number 
 and location of shore stations can be modified as required by the 
 appropriate organization(s). 
 
 For the baseline system the data from the platforms would be trans- 
 mitted to the shore stations at the synoptic hours of 0600, 1200, 1800, 
 and 2400 GMT. These data transmissions can be initiated either by an 
 interrogation transmission from a shore station, or by an onboard clock 
 system. The transmission of a given platform would be in one of the 
 six allocated frequency bands. 
 
 To illustrate the situation, a platform within the primary coverage area 
 of three shore stations is shown in figure 2. The three shore stations 
 can be considered as candidates to receive the data. Thus there are 
 three circuits to consider, one to each of these stations. Furthermore, 
 for each circuit there is a choice of the six frequency bands. Each 
 station-frequency pair will generally have a different predicted relia- 
 bility because the reflection areas,, local times, and noise levels at the 
 shore stations will be different. This results in a total of 18 possible 
 station-frequency combinations. The object, of course, is to select the 
 pair with the highest predicted reliability. 
 
 With 18 shore stations and 6 frequencies there are 108 such combinations 
 on a worldwide basis. This means that a maximum of 108 transmissions 
 could be accommodated at a given time. The actual number would be 
 significantly less than this because many of the station-frequency combina- 
 tions might be unreliable for a given platform; or a particular combination 
 would be likely to interfere with another combination. With a possibility 
 of having several hundred buoys and several thousand instrumented ships, 
 it is necessary (and also technically desirable) to divide each of the six 
 allocated 3. 5 kHz-wide bands into several separate narrower bands 
 (channels). As an example, if each band were to be divided into 13 such 
 in-band channels, each 200 Hz wide*, (leaving 900 Hz for guard space) 
 every station could then receive data from 13 platforms in the same 
 
 * The example of 200 Hz wide channels is purely arbitrary and should 
 not be taken as a recommendation of channel bandwidth for the IGOSS. 
 
CO 
 
 <3 
 +i 
 
 to 
 
 0) 
 Sh 
 O 
 
 CO 
 
 o 
 
 s 
 
 +i 
 
 CO 
 
 CO 
 CO 
 
 o 
 ex. 
 
 
frequency band at the same time. This increases the number of combina- 
 tions to 13 times 108 or, 1404. It is also possible to receive data at 
 different times, but still with the allowable time limit following the 
 synoptic hour. In other words, one buoy might be read out between 1800 
 and 1802 GMT, and another between 1802 and 1804 GMT, etc. These are 
 called timeslots, and for the baseline system, it is assumed that there 
 are six of them for each synoptic hour. The total is now 6 times 1404 
 or 8, 424 station, frequency, in-band channel, and time slot combinations 
 available for gathering data. The object is to assign these combinations 
 (resources) in such a way that the performance of the telemtery system 
 is optimum. The overall telemetry system performance resulting from 
 such assignments is discussed in section 3. 
 
 Figure 2. A sketch showing three possible candidate stations 
 for communications to a single buoy. 
 
 9 
 
3. TELEMETRY SYSTEM PERFORMANCE 
 
 The telemetry system reliability (TSR) at a given synoptic hour can be 
 defined as the expected percentage of data messages from all platforms 
 that is received at the shore stations with acceptable error rates. This 
 reliability is the average of all the individual, predicted circuit relia- 
 bilities. For example, a given assignment of station-frequency-channel- 
 timeslots (SFCT) might produce a TSR of 90 percent with individual 
 circuit reliabilities ranging from 60 percent to 97 percent. 
 
 The telemetry system reliability and the distribution of reliabilities are 
 functions of a number of parameters as discussed in section 1. 2. Some 
 of these can be changed or controlled and others cannot. Parameters 
 we can control are: 
 
 (1) Transmitter RF power output. 
 
 (2) Antenna gains and radiation patterns. 
 
 (3) Required signal-to-noise ratio and susceptibility to multipath. 
 (Various methods of modulation and detection will affect the 
 performance in terms of error rate. Some methods will 
 require higher or lower signal-to-noise ratios for the same 
 error rate, and some will tolerate more or less time spread 
 of the received signal that is inherent in HF propagation). 
 
 (4) Number and location of shore stations. (Increasing the number 
 of shore stations will increase the number of candidate stations 
 and may increase some circuit reliabilities. ) 
 
 (5) Number and distribution of transmitter frequencies on each 
 platform. (For economy, a platform might be equipped to 
 transmit on only two or three frequencies rather than all 
 
 six. At certain times the best frequency may not be available, 
 which decreases reliability. Likewise, the m-band channel 
 may be either fixed or switchable from shore. The latter 
 case would lead to higher reliabilities because interference 
 with transmissions from other platforms could be avoided by 
 changing in-band channels. ) 
 
 (6) Number of in-band channels and timeslots. (Increasing the 
 number of in-band channels per allocated frequency band and 
 increasing the number of timeslots permits greater flexibility 
 
 10 
 
in assigning station-frequency combinations and therby 
 increases the circuit reliabilities. ) 
 
 (7) Assignment method. (If the system capacity is inadequate so 
 that several platforms must compete for a good assignment, the 
 method in which assignments are made will affect the telemetry 
 system reliability. As an extreme example, platforms could 
 be randomly assigned to SFCT combinations without regard to 
 propagation predictions. This would oe simple, but the 
 resulting system reliability would be low. ) 
 
 (8) Geographic distribution of moored buoys. (Conceivably the 
 moored buoys could be placed where propagation conditions 
 are favorable, thereby increasing the associaicu circuit 
 reliabilities. For a given set of shore stations then, the geo- 
 graphic distribution of moored platforms will have an effect 
 on the TSR. ) 
 
 (9) Choice of data transmission time. (For the baseline system 
 we assume that the data are to be sent at the 4 synoptic 
 hours mentioned previously. For a system where this is not 
 
 a requirement, it is possible to send the data when propagation 
 conditions are predicted to be the best. ) 
 
 Parameters we cannot control are: 
 
 (1) Time of year or season and level of solar activity. (Since 
 deploying the platforms is a long-term proposition, circuit 
 reliabilities will vary with season and solar activity. ) 
 
 (2) Geographic distribution of mobile platforms. (This is actually 
 a semi-controllable factor. One can, for example, choose to 
 instrument ships involved in the North Atlantic trade routes and 
 thereby concentrate platforms in that area. However, even in 
 such a case the geographic distribution will change from day to 
 day. This means TSR will change with the day-to-day changes 
 in ship locations. ) 
 
 These uncontrollable factors make simulation necessary. A "good" 
 system design is one in which the TSR is optimized by a choice of the 
 controllable parameters for the expected range of the uncontrollable 
 parameters. 
 
 11 
 
In the selection of the designed controlled variables, it is apparent that 
 there are many trade-offs. The cost of a 3 dB increase in transmitter 
 power can be compared with the cost of a 3 dB increase in antenna gain, 
 or the cost of a more complex modem that would require 3 dB less SNR 
 for the same error rate. From a systems standpoint, each would 
 result in the same increase in TSR. This same increase in TSR might 
 also be accomplished by an increase in the number of shore stations or 
 by an increase in the number of channels or timeslots. The simulation 
 program, which is described in the next section, investigates the TSR 
 as a function of these parameters. In that way, the trade-offs can be 
 evaluated in terms of cost/ effectiveness. 
 
 4. PROGRAM DESCRIPTION 
 
 The computer based simulation model depends upon the HFMUFES 
 computer program that was mentioned earlier for predicting HF 
 communication performance. The program has been extensively 
 modified, and most general purpose features have been eliminated. 
 The basic propagation model has not been changed. 
 
 A block diagram of the simulation model is shown in figure 3. The main 
 computer program is called SIMIGOSS (for SIMulation and IGOSS). It in 
 turn has three main subroutines: (1) OPTIMAL, which makes the actual 
 assignments, (2) PREDPRFM*, the basic circuit performance pre- 
 formance prediction routine, and (3) SKEDULE, the statistical generation 
 and assignment output subroutine. 
 
 In the computer simulation model, SIMIGOSS, predicting individual 
 circuit reliability of each of the six frequencies and 4 synoptic hours 
 now takes about 7 sec of computer time. Even at this rate, it is too 
 expensive to re-run the simulation model program for each platform/ 
 shore station pair whenever controllable parameters are changed, or a 
 different geographic distribution of platforms is assumed. Therefore, 
 for this simulation model the ocean areas of the world have been divided 
 into approximately 3, 000 areas for which the performance of any circuit 
 originating within each area can be considered a constant for a given 
 shore station and set of input parameters. 
 
 * PREDPRFM utilizes a number of HFMUFES subroutines. For a brief 
 description of these subroutines (as given by Barghausen et al. ) see 
 appendix A. 
 
 12 
 
Shore Station 
 Locations and 
 Control Data 
 
 SIMIGOSS 
 
 Predictions 
 
 OPTIMAL 
 
 SKEDULE 
 
 Assignments (Station, Frequency, 
 Channel, Timeslots ) 
 
 Platform Data 
 By Area 
 
 Ionospheric and Noise 
 Predictions 
 
 PREDPRFM 
 
 m Summaries (TSR, Histograms of 
 Reliability, Resource, Utilization, 
 Signal Distributions) 
 
 Figure 3. Block diagram of the simulation computer model. 
 
 Propagation predictions for these areas are then computed, and the 
 results (such as path loss and probability of ionospheric support) are 
 stored on magnetic tape. Changes in the geographic distribution of 
 platforms, transmitter power, antenna gains, required SNR, or any- 
 other designer specified parameter (except shore station location) can 
 then be evaluated in terms of TSR without re-running the propagation 
 prediction portion of the program. 
 
 The inputs to SIMIGOSS follow: 
 
 (1) Requirements by prediction area. (This includes the number of 
 platforms in the area, frequencies and channels associated 
 with each platform, and a list of the candidate shore stations. 
 Limiting stations to the three closest rather than all 18 decreases 
 computation time. Other criteria, besides the closest three, 
 
 for assigning candidate stations can be incorporated easily. ) 
 
 (2) Ionospheric and noise parameters. (This is the basic data 
 upon which propagation predictions for individual circuits 
 are based.) 
 
 13 
 
(3) Shore station locations and other data. (The data includes 
 controllable system characteristics, such as transmitter 
 power, and the month and level of solar activity. The program 
 is quite flexible with respect to these parameters. ) 
 
 (4) Circuit predictions by area. (The predictions are stored on 
 magnetic tape and can be used either on input or on output. If 
 the predictions from the area to the candidate stations have 
 already been calculated for the month and level of solar activity, 
 predictions are read from the tape. If not, they are computed 
 and stored on tape for future re -use. ) 
 
 The outputs of the program are: 
 
 (1) Assignments. (That is, each platform for each synoptic hour 
 is assigned to transmit its data to a specified shore station on 
 a specified frequency, channel, and timeslot. ) 
 
 (2) Predictions. See (4) above. 
 
 (3) Summaries for analysis. (These include the telemetry system 
 reliability for each synoptic hour, histograms of circuit 
 reliabilities, resource utilization, interference /conflict 
 problems, and distributions of signal strength, antenna takeoff 
 angles, and antenna bearings, that result from the assignments 
 made. As currently implemented, the statistical summaries 
 other than TSR are obtained from a separate program that will 
 eventually be incorporated in the SKEDULE subroutine. ) 
 
 The operation of the program is as follows: 
 
 (1) The control information and prediction data are read in and 
 stored. 
 
 (2) In the current assignment technique, the data for the first (or 
 next) area are read from tape and stored in the computer. 
 
 (3) If predictions of reliability have not been previously obtained 
 for the area, the PREDPRFM is called and circuit reliability 
 predictions are computed for the circuit from this platform 
 to the candidate stations for the six frequencies and the 4 
 synoptic hours. The best station-frequency that has an un- 
 assigned timeslot on the channel associated with the platform 
 is then assigned. 
 
 14 
 
(4) Each platform in the area is then considered in turn. This is 
 done for each synoptic hour. If the timeslots are all exhausted 
 for the best station-frequency, the next best station-frequency 
 combination is considered, and so forth. 
 
 (5) When the station-frequency-channel-timeslot combination has 
 been assigned, the assignment and the relevant data are 
 recorded on magnetic tape, and (at the option of the operator) 
 printed. The assignment data are also accumulated for 
 statistical purposes. If no unassigned SFCT combination can 
 be found, the operator is informed in the printout. 
 
 (6) Steps are repeated until all platforms have been assigned to a 
 shore station, or the platform is noted as unassignable. 
 
 (7) Statistical results are then presented. 
 
 Before going to an illustration of a typical simulation run some limitations 
 of the current model should be noted. 
 
 (1) Interference is not considered except that a station cannot be 
 assigned to receive more than one platform on the same 
 frequency, channel, and timeslot. In effect this allows unlimited 
 geographic sharing. 
 
 (2) Multipath effects are not considered in the assignment. Multi- 
 path (the arrival of the signal via more than one path so that the 
 signal is "smeared" in time) may seriously degrade the circuit 
 performance on a given frequency. 
 
 (3) Groundwave paths are not considered. When a platform is 
 located very near a shore station, it may be possible to receive 
 the data via groundwave, and the groundwave signal may inter- 
 fere with the reception of the skywave signal. 
 
 (4) A "sequential" assignment technique is used. With the 
 assumption of 6 timeslots, 13 channels per frequency, and 18 
 shore stations, conflicts begin to occur in later assignments 
 when a thousand or so platforms are considered. This means 
 that the SFCT resources assigned early might be used more 
 advantageously in the later assignments, thus implying that 
 there may be another distribution of resources that would lead 
 to a higher telemetry system reliability. This is a result of 
 sequential assignment technique not always leading to an 
 
 15 
 
optimal assignment. If the areas are processed sequentially 
 but are sorted first in order of importance, this method may- 
 have merit. Otherwise, if the capacity of the system is taxed, 
 a more sophisticated assignment technique may be needed, or 
 the capacity must be increased by increasing the number of 
 timeslots. The trade-off here is that for a given timeslot 
 length, an increase in the number of timeslots will eventually 
 exceed the allowable transmission period associated with each 
 synoptic hour. Data received after this may be less useful 
 for severe weather or ocean condition detection, or forecasting 
 purposes. Also, the amount of time available outside the 
 interrogation period would be reduced. The competition for 
 favorable assignments will be even more severe when inter- 
 ference is considered fully. 
 
 (5) Sporadic E modes are not considered. 
 
 5. SAMPLE RESULTS 
 
 This section of the report presents a sample of the outputs from the 
 SIMIGOSS program. For this sample, computations were made using 
 the simulation model with a worldwide distribution of 5, 500 platforms. 
 One platform was located in each of the 3, 313 prediction areas, and 
 the remaining 2, 187 platforms were then distributed randomly over all 
 areas. The total number of platforms and their geographic distribution 
 is not realistic, but serves to exercise fully the SIMIGOSS program. 
 Data on world ship traffic patterns is currently being analyzed to provide 
 more realistic distributions for future runs. The other input data and 
 assumptions for the sample computations with the simulation model were 
 as follows: 
 
 (1) June was chosen as representative of a Northern Hemisphere 
 summer, or Southern Hemisphere winter. 
 
 (2) The smoothed sunspot number (SSN) is assumed to be 110 for 
 the month, which represents a high solar activity. 
 
 (3) The platform transmitter is assumed to deliver 100 W of power 
 to the platform antenna on all frequencies. 
 
 (4) The platform antenna is assumed to be a discone, 11m high, 
 with a maximum power gain of 6 dB. The shore station antenna 
 is assumed to be a vertical antenna with a 12 dB power gain. 
 
 16 
 
(5) The required SNR for a 10"^ error rate is assumed to be 38 dB 
 for a FSK modulation system with diversity reception. 
 
 (6) The man-made radio noise at shore stations is assumed to be 
 typical of an urban environment. 
 
 (7) The shore stations are those shown on the world map in 
 figure 1. 
 
 (8) Multipath, sporadic E, and groundwave propagation are not 
 considered in the predictions. 
 
 A sample assignment of platforms to shore stations for one prediction 
 area is shown in figure 4. The top line shows the area number (103), 
 the geographic coordinates (6. 36°S, 6. 40°W), and the number of plat- 
 forms in the area (2). The remainder of the page is divided for the 4 
 synoptic hours. At the 0600 GMT synoptic hour, platform number 168 
 has been assigned to station number 10 (Nairobi), on frequency band 
 number 2 (6 MHz), on channel 3 and in timeslot number 1. Predicted 
 reliability for this circuit is 98 percent and the predicted signal strength 
 at the shore station is -73 dBW. The latitude and longitude of the shore 
 station is 1. 17°S and 36. 50°E. The distance from the area to the station 
 is 3, 387 km. 
 
 The TSRs for each of the 4 synoptic hours are: 
 
 0600 GMT 80 percent 
 
 1200 GMT 78 percent 
 
 1800 GMT 78 percent 
 
 2400 GMT 73 percent 
 
 A histogram showing the percentage of the total assignments in reliability 
 ranges of 10 percent is shown in figure 5. Figures showing TSR, and 
 histograms of this type are useful outputs for system design. A histo- 
 gram that shows the percentage of the assignments that were made to 
 the best station-frequency pair, the second best, and so on (as a result 
 of the limited number of channel/timeslots) is presented in figure 6. 
 This figure is useful in studying the communication system capacity. A 
 histogram showing station utilization is shown in figure 7. This 
 information is operationally useful in showing the relative workload at 
 each station. It is also useful in evaluating shore station locations during 
 the design phase. Figure 8 is a histogram illustrating frequency utilization, 
 and is useful in indicating the frequencies that will be most useful on the 
 platforms. 
 
 17 
 
AREA NO. 103 ApEa LATITUDE = -6.36 AREA LONGITUDE = -6.40 NO. OF PLATFORMS IN THE AREA 
 
 ^* + + uu-w*ui^Hi*ii + 4Kiii^^ti + ^^^;ii^HAi^ii4i + i + mi + + ^ HOUR 
 
 PLATFORM AREA STATION FRLQ CHANNEL TiMFSLOT RFLlAB SIGNAL 
 
 NO. NO. NO. NO, NO. PERCENT -DBW 
 
 167 103 10 Z 3 1 98 73.0 
 
 168 103 10 2 13 1 98 73.0 
 
 * + ^** + ^^4>** + 4444<<l'4'<W*4-*4-4^ + 4'4-4-**^4'**<l4'^14'4.4*4^4'4'4'4'4'44.4'44-* HOUR 
 
 PLATFORM AREA STATION FRL'J CHANNEL TjMESLOT RELlAB SIGNAL 
 
 NO. NO. NO. NO. NO, PERCENT -DB* 
 
 167 103 11 6 9 1 94 88.0 
 
 168 103- 11 6 10 2 9* 88.0 
 
 + ii»^* + *i+w*i;il+4l^ii*u^4 + + **i^^+^l^l*iiilii*i + il^+ + l+^l HOUR a 180n GMT 4- 4. 4. 4- 4- 4. 4- 4- * 4. 4- 4- 4 4- 4- 4- 4- 4 4. 4 4. 4. 4. 4 4 4 4 
 
 platform area station rRto channel timeslot reliab signal station station oist 
 
 NO. 110. NO. NO. NO. PERCENT -DBW LATITUOE LONGITUDE KM 
 
 167 103 8 4 13 4 97 8i.O 14.38 -17.27 3483 
 
 168 103 8 4 11 2 97 Bl.O 14.38 -17.27 3483 
 
 + + J. + ** + 4,4'**4. + + J' + **4.J.J.*4'*4.4.4,4. + J,*4' + + 4<*^4,*4,4.4.4.4,4. + + '1.4.4,4. + *4.-t'4. + J'+ HOUR = 240r> GMT 4, 4, 4, + 4. 4, 4. 4, 4/ 4- 4 4. 4. 4- 4, 4, 4, 4, 4/ 4. 4, 4. 4. 4. 4. 4, 4. 
 
 PLATFORM AREA STATION FRtQ CHANNEL TIMESLOT RELIAB SIGNAL STATION STATION OlsT 
 
 NO. NO. NO. NO. NO. PERCENT -03* LATITUOE LONGITUDE KM 
 
 167 103 8 3 12 1 97 76.0 14.38 -17.27 3483 
 
 168 103 8 3 8 2 97 76.0 14.38 -17.27 3483 
 
 60" GMT 
 
 4,4/4,4.4.4,4.4.4.4.4.4,4.4.4.4.4.4.4.4.4.4.4.4/4.4.4. 
 
 STATION 
 
 STATION 
 
 DIST 
 
 LATITUOE 
 
 LONGITUDE 
 
 KM 
 
 -1.17 
 
 36.50 
 
 3387 
 
 -1.17 
 
 36.50 
 
 3387 
 
 120n GMT 
 
 4.4.4.4,4.4.4.4.4,4.4.4.4.4.4.4.4.4.4.4.4.4.4/4/4,4.4. 
 
 STATION 
 
 STATION 
 
 DIST 
 
 LATITUOE 
 
 LONGITUDE 
 
 KM 
 
 -25.45 
 
 28.12 
 
 3135 
 
 -25.45 
 
 28.12 
 
 3135 
 
 Figure 4. A sample output from the assignment routine . 
 
 Many other outputs from the program are possible, but the ones just 
 illustrated are the most useful in the current stage of the system design. 
 An example of another output, which has not yet been programmed, 
 would be a summary of antenna take-off angles for each frequency, 
 and azimuthal bearings to assigned platforms, so that the antenna 
 requirements for each shore station can be determined. This would 
 provide a basis for antenna design. 
 
 Normally several computations using different constraints on the simulation 
 model would be made to test effects of changing one or more of the design 
 parameters. For example, TSR can be plotted as a function of both buoy 
 transmitter power and the number of timeslots (reflecting the system 
 capacity) of the system. For examples of such runs, the reader is 
 referred to the report by Adams et al. (1969). 
 
 The results given in this section should be considered examples only, 
 since (1) they consider rather unrealistic numbers and distributions of 
 platforms, (2) no attempt has been made at optimizing the system or 
 assignment method, and (3) potential interference between operations 
 on the same frequency at the same time is not considered. 
 
 18 
 
1.0 
 
 ■D 
 
 C 
 
 < 
 
 -Q 
 O 
 
 a; 
 
 q: 
 
 E 0.5 
 o 
 
 o 
 
 CL 
 
 O 
 CL 
 
 c 
 
 Q) 
 O 
 
 
 
 —*- h T ^ 1 [ PH^lfl I 
 
 1 ' 
 
 10 20 30 40 50 60 70 80 9 
 
 Reliability in Percent 
 
 100 
 
 Figure 5. A histogram showing the distribution of shore 
 platform assignment circuit reliabilities . 
 
 19 
 
10000 
 
 8000 ma 
 
 E 
 
 CL 
 
 o> 
 
 E 
 
 3 
 
 6000 
 
 4000 
 
 2000 
 
 
 
 I 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 
 
 Order of Choice 
 
 Figure 6.. A histogram showing the number of platform- 
 stations assignments made as a function of circuit 
 reliability choice. 
 
 20 
 
2500 
 
 C 
 
 o 
 
 o 
 CO 
 
 u 
 
 o 
 UJ 
 
 <D 
 
 C 
 
 < 
 
 E 
 
 CL 
 
 -O 
 
 E 
 
 13 
 
 2000 - 
 
 1500 - 
 
 1000 - 
 
 500 - 
 
 
 
 2 3 4 5 6 7 8 9 10 ' II 12 13 14 15 16 17 18 
 
 Each Number on the Abcissa Identifies a Receiving Station 
 
 Figure 7. A histogram showing the number of platforms 
 assigned to each shore station. 
 
 21 
 
0.5 
 
 ■o 
 o> 
 
 c 
 o> 
 
 "</> 
 a) 
 
 < 
 
 u 
 
 c 
 o> 
 
 CT 
 
 <V 
 
 0.4 
 
 0.3 
 
 o 
 
 c 
 Q 
 
 0.2 
 
 0. 
 
 
 
 **W*»WWfW 
 
 If 
 
 il 
 
 4.16 
 
 111 
 
 I 1 I IliiM 
 
 6.25 8.33 12.48 16.34 
 Frequency in MHz 
 
 22.16 
 
 Figure 8. A histogram showing the distribution of station- 
 •platform assignments as a function of the allocated 
 frequencies . 
 
 22 
 
6. SUMMARY AND CONCLUSIONS 
 
 (1) There are marked variations in sky wave propagation depending 
 upon path length, geographic location and especially upon time 
 of operation. 
 
 (2) For a given path and time it is essential to choose the proper 
 frequency for successful data collection via sky waves. 
 
 (3) Since the major variations in sky wave propagation are predictable 
 many years in advance, it is possible to use these predictions in 
 the design of efficient data collection systems including the 
 frequency assignment and time scheduling of the data collection. 
 
 (4) The simulation of proposed data collection systems by computers 
 permits an evaluation of these systems. This evaluation can be 
 fast and efficient. 
 
 (5) It appears desirable to plan a worldwide system complete with 
 frequency planning rather than starting many small systems and 
 attempting to integrate additional circuits on a piecemeal basis. 
 
 (6) The model illustrated in the report indicates the feasibility of 
 data collection in the presence of natural noise limitations. When 
 the model is extended to include the effect of mutual interference 
 between circuits which may be operating co-channel or adjacent 
 channel it would form the basis for international coordination of 
 
 a worldwide oceanographic data collection system. 
 
 7. ACKNOWLEDGMENTS 
 
 The authors acknowledge the contribution of Mr. Alfred F. Barghausen 
 and Mr. James W. Finney in providing the long-term prediction routine. 
 The help of Mr. Judd A. Payne in programming for the computer model 
 is also appreciated. Mr. Ted deHaas contributed greatly to the material 
 in section 1. 1. 
 
 23 
 
8. REFERENCES 
 
 Adams, J. E. , D. N. Hatfield, and J. A. Payne (1969), A feasibility 
 study of a proposed national oceanographic data buoy HF com- 
 munication system, unpublished 1 ESSA Tech. Memo, ERLTM- 
 ITS-205 
 
 Barghausen, A. F. , J. W. Finney, L. L. Proctor, and L. D. Schultz 
 (1969), Predicting long-term operational parameters of high- 
 frequency sky-wave telecommunication systems, ESSA Tech. 
 Rept. ERL-110 ITS 78 (U. S. Government Printing Office, 
 Washington, D. C, 10402). 
 
 Haydon, G. W. (1967), Theoretical evaluation of band 7 frequency 
 
 complements for ocean data communications, unpublished 1 ESSA 
 Tech. Memo, IERTM-ITSA-54. 
 
 Jones, W. B. , R. P. Graham, and M. Leftin (1966), Advances in iono- 
 spheric mapping by numerical mapping, NBS Tech. Note 337 
 (U. S. Government Printing Office, Washington, D. C. 20402). 
 
 Lucas, D. L. and G. W. Haydon (1966), Predicting statistical performance 
 indices for high frequency ionospheric telecommunications systems, 
 ESSA Tech. Rept. IER1-ITSA 1 (U.S. Government Printing 
 Office, Washington, D. C. 20402). 
 
 This document is in the public domain, but it is considered unpublished 
 since it was not printed for wide public distribution. 
 
 24 
 
APPENDIX A 
 Description of Subroutines Used by PREDPRFM 
 
 The subroutine PREDPRFM corresponds to the main program portion 
 of HFMUFES. The subroutines and their descriptions, which are 
 used by PREDPRFM, are given here. The descriptions are largely 
 taken from Barghausen et al. (1969) where more detailed information 
 is available. 
 
 VERSY 
 
 MAGFIN 
 
 BEMUF 
 
 LUFFY 
 
 RELBIL 
 
 NOISY 
 
 GENFAM - 
 
 evaluates coefficients for the world maps of iono- 
 spheric parameters that have been generated as a 
 function of Universal Time (UT) and latitude (geographic 
 or magnetic dip). 
 
 computes the magnetic field components of the earth 
 at any height and geographic latitude and longitude. 
 
 calculates, on the basis of parabolic layer theory, the 
 maximum usable frequency, takeoff angle, and virtual 
 height of reflection, or, for a specific frequency, only 
 the last two parameters. 
 
 calculates operational parameters for up to seven 
 modes for each hour and frequency. Uses RELBIL 
 to calculate SNR and reliabilities. (The original 
 LUFFY in HFMUFES has been significantly modified 
 to eliminate features not needed in the SIMIGOSS 
 program. 
 
 calculates atmospheric, galactic, and man-made noise 
 and then determines controlling noise. Also calculates 
 the SNR and reliability. 
 
 evaluates the world maps of noise and land areas that 
 have been mapped by a Fourier series. 
 
 calculates frequency dependence of the median 1 MHz 
 atmospheric noise and its associated decile values and 
 standard deviation of the median and upper and lower 
 deciles of the atmospheric noise. 
 
 25 
 
CHISQ 
 
 evaluates the chi-square probability function. 
 
 SYSSY 
 
 obtains from a table the values of median excess 
 system loss, upper and lower standard deviations, 
 and the error in predicting excess system loss and 
 upper and lower standard deviations in the error. 
 
 GAIN 
 
 calculates gain in decibels of the transmitter and 
 receiver antennas. (In HFMUFES this is a very 
 general subroutine. It has been completely changed 
 for SIMIGOSS to compute gains as a function of 
 frequency, takeoff angle, and bearing for only 
 antennas contemplated for IGOSS. ) 
 
 GPO 859 -581 
 
 26 
 
PENN STATE UNIVERSITY LIBRARIES 
 
 A0Q0Q72D2DQ7M