An examination of elevated frequency propagation over a transpolar path

Transcription

1 RADIO SCIENCE, VOL. 39,, doi: /2002rs002850, 2004 An examination of elevated frequency propagation over a transpolar path John M. Goodman Radio Propagation Services, Inc., Alexandria, Virginia, USA John W. Ballard Radio Propagation Services Inc., Los Altos, California, USA Received 29 November 2002; revised 4 April 2003; accepted 2 June 2003; published 5 February [1] A Chirpsounder 1 system has been installed to evaluate the communication path between Svalbard (Spitsbergen) and Barrow, Alaska. This system of measurements is used to update operational ionospheric propagation models we use to nowcast the effectiveness of high-frequency (HF) communications over polar paths. Operations have been virtually continuous since December 2000, during which time Radio Propagation Services Inc. has made measurements of the propagation effects that have been encountered. Oblique sounder patterns have been diverse, ranging from standard (i.e., classical) ionograms with a sharp nose near the anticipated maximum useable frequency to rather diffuse patterns. However, the most interesting patterns consist of nonclimatological patterns of signals associated with EMOFs. These were first observed to occur during the winter, but they are in evidence throughout the year. The strong signal strengths associated with these EMOF patterns are transient and consistent with independent ionospheric layers or modulations of existing layers. The EMOF signals have been analyzed statistically, and the paper will describe the diurnal and seasonal behaviors. We will also investigate the association of these observations with polar patches and blobs, which have been observed to convect across the polar cap, especially in the winter. These observations are not simply scientific curiosities. The character and persistence of these signals have a significant impact on communication effectiveness for over-the-pole air transport operations. Issues of radio spectrum, service provider location, and dynamic frequency management are directly related to the phenomena observed. INDEX TERMS: 2439 Ionosphere: Ionospheric irregularities; 2475 Ionosphere: Polar cap ionosphere; 2487 Ionosphere: Wave propagation (6934); 2494 Ionosphere: Instruments and techniques; KEYWORDS: ionosphere, polar patches, Chirpsounding, HF propagation Citation: Goodman, J. M., and J. W. Ballard (2004), An examination of elevated frequency propagation over a transpolar path, Radio Sci., 39,, doi: /2002rs Background [2] Increasing emphasis is being placed on the use of shortwave communications (e.g., 3 30 MHz) to support the commercial aviation community. The ARINC GLOBALink/HF system for reliable data link service has been initiated in the year 2000, and this system is a worldwide HF communication system serving commercial aviation for oceanic transits. Communications at Copyright 2004 by the American Geophysical Union /04/2002RS HF are known to present a challenge because of strong ionospheric interactions. Various advanced modems and forms of automatic link establishment have been advanced to assist in the process of link connectivity when viable propagation paths actually exist. HF communication requirements in Antarctica have been established to support NSF-sponsored science missions, and the McMurdo HF modernization program addresses some deficiencies, notably with antennas. Moreover, in the Arctic region, commercial flights over the pole may have an increasing significance in the coming years, and this will make exploitation of technology for mitigation of deleterious propagation effects all the more essential. The U.S. government provides various 1of12

2 categories of solar-terrestrial data via the Web, and these data can be exploited to provide guidance to communication specialists and air traffic controllers in the context of high-latitude communication reliability. Radio Propagation Services Inc. (RPSI) has developed a unique capability to take advantage of this data as well as other resources under its control. Using its own sounder data and proprietary processing algorithms, RPSI has developed tools for nowcasting and forecasting the performance of HF links between aircraft and high-latitude ground stations in the context of over-thepole operations. This paper sketches the arena of highlatitude propagation and cites previous data, models, and studies. We extrapolate from a comprehensive northern experiment to draw inferences about what can be achieved under pathological conditions. RPSI has also taken data over a specific high-latitude path (e.g., Svalbard to Barrow). [3] The National Science Foundation requires timely and reliable communication within the Antarctic region, including the South Pole, as well as long-haul connectivity from McMurdo to stations such as Christchurch, New Zealand. It is noteworthy that there is little alternative to shortwave communications within Antarctica owing to the lack of satellite coverage and an alternative communication infrastructure. However, shortwave or HF communication, while available for Arctic and Antarctic aeronautical communication, may suffer from outages as a result of magnetic storms and solar particle events. These events are the space weather analogues to well-known atmospheric weather events, and they are characterized by the properties of the Arctic ionosphere. The ionosphere is the medium used to establish longhaul HF communications for transpolar flights. In these situations, there are two ways to solve the problems associated with HF communications. The first way is to engineer the solution using the most robust system technology, including adaptive modems with equalization, and with the incorporation of the most appropriate frequency and station diversity possible. However, this cannot stand alone. The second way, fully integrated with the first, is to apply reliable nowcasting and forecasting technologies that will enable real time frequency management and station selection criteria to be invoked for mitigation or circumvention of troublesome propagation effects. Using its patented sounder-update technology, RPSI has advanced a scheme for achieving optimal HF communication in the Arctic [Goodman and Ballard, 1999a]. 2. High-Latitude Propagation Experience [4] RPSI is currently supporting the operational frequency management mission of GLOBALink/HF. This system, operated and managed by ARINC, is a worldwide HF data link communication system (HFDL) for commercial carriers. The HFDL network is composed of a number of stations, including several used to cover Arctic areas. These include Barrow, Alaska, and Reykjavik, Iceland. RPSI uses its Dynacast 1 technology to develop weekly and emergency active frequency tables (AFTs) that are used to accommodate successful connectively and reliable communications at high latitudes [Goodman and Ballard, 1999b]. The company is also involved in several research and development (R&D) activities that are directed toward an improved understanding of polar ionospheric conditions and Arctic communication issues. [5] A major R&D activity was launched in December 2000 and continued through much of We have deployed a Chirpsounder 1 transmitter in Svalbard (Spitsbergen) and a receiver in Barrow, Alaska. With this system we have observed nonclimatological patterns of signal scatter at EMOFs, especially in the local wintertime. These propagation disturbance episodes are likely associated with blobs, patches, or tongues of ionization. The ionization features, collectively called patches for convenience herein, have been noted for some time. Our EMOF observations, using the obliqueincidence sounding technique, are likely symptomatic of the same phenomena detected using vertical-incidence sounders and incoherent scatter radars. SuperDARN radars may also have detected these phenomena. Indeed, patches of F region plasma, having horizontal dimensions of km, are a commonplace feature of the polar ionosphere. It is thought that these patches, having elevated electron densities, are entrained within the convection pattern across the polar cap in the antisunward direction. The process appears to be correlated with enhanced magnetic activity when the interplanetary magnetic field (IMF) is directed southward and during the winter when the polar region is in quasi-darkness. However, there is a wide range of characteristics. For example, when the IMF is northward, electron density enhancements in the Sun-Earth direction can also occur but with reduced amplitude. Such features would be expected to produce terrestrial propagation effects, and we anticipate increased ionospheric support whenever a patch with elevated values of electron density traverses a communication path. Using our oblique Chirpsounder system, we have detected these EMOF events. Moreover, judging from the duration of the disturbances, we find them to be relatively large-scale features. [6] The RPSI oblique sounder studies have detected EMOF events within the north polar cap region, and the properties are consistent with our understanding of polar patches and their behavior. The understanding of these phenomena should have relevance to Antarctica as well. Later on we shall comment on the use of these effects in 2of12

3 Figure 1. Depiction of the HF ray trajectory from Svalbard (Spitsbergen) to Barrow (Alaska). The Chirpsounder 1 receiver is located in Barrow, from which the data are uploaded to a Web server located in Los Altos, California. the context of spectrum planning and frequency management in other areas of the report. 3. General Description of Plasma Blobs and Patches [7] Buchau et al. [1983] have examined the wintertime polar cap F region with an emphasis on its structure and dynamics. Crowley [1996] has reviewed ionospheric patches and blobs and discusses the distinctions between the two. Discrete electron density enhancements occurring in the F region of the ionosphere at high latitudes, with horizontal scales of km, are called patches if inside the polar cap (or entering), and they are called blobs if they have passed through the cap and are outside of it. These plasma enhancements are a factor of 2 or more above ambient, and steep gradients associated with patches and blobs are correlated with smallscale structure (i.e., irregularities). [8] The status of research about patches provided by Crowley [1996] has been augmented by Rodger [1998, 1999], who describes polar patches as regions of high F region plasma concentration observed in the polar cap. One of the critical issues remaining is the differentiation in the observations between the distorted tongue of ionization that may convect across the polar cap from genuine polar patches which Rodger describes as isolated islands of high concentration. 3of12 [9] Terminology for the phenomena has been given in various papers [Crowley, 1996; Rodger, 1999; Benson and Grebowski, 1999]. The latter authors describe the polar cap as the region poleward of the auroral oval in each hemisphere. In the winter it is subject to prolonged periods of darkness and has a lower electron density (Ne) than in the oval but contains both Ne enhancements (i.e., tongues, patches, and blobs) and Ne depletions (namely troughs, cavities, and holes). In its simplest conception, ignoring more complicated structures, the average largescale polar cap Ne cavity deepens during the winter night as plasma convects and chemically decays in the absence of a well-defined source of photoionization. [10] There is an intriguing result from stations in Vostok, Mirny, and Dixon that shows a high correlation between noontime fof2 values (for wintertime solar maximum conditions) and dynamic pressure of the solar wind. Rodger [1999] speculates that the ionospheric convection pattern is quite expanded when the solar wind velocity is high. He reckons that lower-latitude plasma is drawn into (i.e., entrained) the cross-polar tongue of ionization and hence over the observatories used in the study. [11] There are a number of source materials that may be drawn upon. Kelley [1989] has written an impressive book on ionospheric physics. It covers essential theory about high-latitude electrodynamics, current systems, convection patterns, and irregularities. Benson and Grebowski [1999] have found that low ionization heights are sometimes observed in topside sounder data. They feel that these lowered heights may have been misinterpreted (by others) as F region holes since the peak layer heights drop monotonically during disturbed times. The implication of topside sounder results on highly oblique paths is not totally evident. However, it is clear that if Ne is fixed and the layer height were to drop dramatically, then the maximum observable frequency (MOF) exploiting that disturbed layer would increase in proportion to the increase in ray zenith angle, until such time as the height lowering leads to mode switching (i.e., 1-hop to 2-hop conditions). 4. Arctic Radio Propagation [12] There have been numerous studies of HF propagation effects and communication capability within the Arctic environment. A review of these effects is provided in various books and manuscripts [Leid, 1967; Goodman, 1991; Davies, 1990]. Much of our knowledge of the region has been provided by data obtained in the 1950s through the 1970s. [13] From a morphological point of view the highlatitude region is the most interesting part of the ionosphere. It has been said that the auroral zone and associated circumpolar features are our windows to the

4 Figure 2. VOACAP run for 1 December 2000 for a sunspot number of 140. The path is Svalbard to Barrow. The ionospheric coefficients used were the newest URSI values. We have also assumed no contribution to the MUF from sporadic E formations. Use of CCIR coefficients and allowance for sporadic E tend to elevate the MUF curve by several MHz (G. Lane, private communication, 2003). A synthetic oblique incidence sounder record can be determined from use of method 25 in VOACAP using Lane s corrections. This kind of treatment yields a MUF of 20 MHz at 1100 UTC rather than 18 MHz as shown in the graph. See color version of this figure in the HTML. distant magnetosphere, and the presence of visible aurora has fascinated observers for centuries. The interplanetary magnetic field, which may be traced to its solar origins, has a significant impact on the geomorphology of the high-latitude ionosphere and its dynamics, including magnetic substorm development. [14] A wide range of phenomena characterizes the highlatitude region, and the subject is far too complex to be given proper coverage in this abbreviated paper. These phenomena are largely orchestrated by magnetospheric and interplanetary events (of a corpuscular nature) rather than solar (electromagnetic) flux variations. Hunsucker [1983] has examined the salient features with particular 4of12 emphasis on the high-latitude trough. Bishop et al. [1989] have placed these features in a worldwide context. It is well known that the magnetic pole is tilted toward the American sector. This makes high-latitude phenomena more pronounced than would be expected on the basis of geographic coordinates. This is important as we consider HF communication for transpolar flights to/from the United States. 5. Earlier Communication Studies [15] Studies of the polar ionosphere have largely emphasized the ionospheric physics and were less di-

5 Figure 3. We expected ionograms to look like the one shown here. They exhibit a high-ray/lowray pattern, typical of 1-hop F2 layer propagation. In addition, see that the so-called MOF (shown by the dotted line in the lower curve), or maximum observable frequency, is around 14.5 MHz. This is within the margin of error of propagation models in use today (i.e., VOACAP and ICEPAC). We observe that this pattern is not always observed. The estimated MOF, as depicted by the vertical line, is derived from the Dynacast 1 scaling algorithm. rected toward an understanding of HF communications. While ionospheric data were obtained during the international geophysical year (i.e., 1957) and other years leading to the development of useful performance prediction models, many operational issues associated with actual communications were not adequately addressed. One of the more useful forums for examination of practical communications was the 8th meeting of the NATO-AGARD Ionospheric Research Committee in Athens in 1963 [Landmark, 1964]. While this was long ago, a number of Arctic communications phenomena were clearly identified, and more fulsome reports have not been forthcoming since that time. An especially noteworthy paper was presented by Jull [1964]. [16] Jull [1964] examined the results from several studies of oblique-incidence sounding and 30 MHz Riometer absorption with the express objective of using the results to increase the reliability of HF communications over auroral zone circuits. Studies were carried out during high- and low-intensity polar cap absorption Figure 4. An example of an observed ionogram. Ionograms often depart from the example shown in Figure 3. This would be expected since the ionosphere has an intrinsic variability. However, at times the variation is extraordinary. Below it is observed that an ionogram may exhibit additional modes of propagation well above the climatological MUF. Bear in mid that the MUF at the time of observation is of the order of 15 MHz. A prediction system based purely on a climatological model (without updates) would be grossly in error. 5of12

6 Figure 5. Autoscaled estimates of the MOF for 1 6 December This was a period shortly after RPSI began observations from Barrow of signal transmitted from Svalbard. Large diurnal variations were observed. These far exceeded the median values of the MOF (e.g., the MUFs) that are derived from standard models such as ICEPAC and VOACAP. These values are in the MHz range. (PCA) events, and moderate and severe magnetic storms, to determine the effect of these disturbances of HF communications. These paths included point-to-point circuits and on a circuit from a ground station to an aircraft flown into, and through, the auroral zone and into the polar cap. A key finding made by Jull was that frequency sounding systems can be used to provide direct assistance to communication systems in the selection of optimum communication routes and operating frequencies during PCA events and geomagnetic storms. Moreover, in order to match the highly variable characteristics of auroral Es and F layer propagation during geomagnetic storms, it was necessary to provide a wellspaced set of frequency assignments. Finally, because of the rapidity of propagation variability, it was necessary to sample propagation conditions at time intervals no greater than 15 min. Jull reports rather marked differences between model predictions and the optimum working frequencies during transauroral operations, with sporadic E often providing significant connectivity at frequencies both above and below the classical predicted MUF for the designated circuit. Abnormal propagation at elevated HF bands caused by Es patches was shown to be useful during PCA events when the lowest frequencies are not usable. Jull presented some alternate routing strategies that are not wholly pertinent to the current problem, but it Figure 6. Autoscaled estimates of the MOF for 6 28 February This is 23 days. Notice the difference in timescale from Figure 3. Diurnal variations are apparent, especially in the first part of this figure. Again the MOF excursions are far in excess of those predicted in VOACAP or ICEPAC. 6of12

7 Figure 7. Autoscaled estimates of the MOF for July The general pattern of MOFs is more consistent with long-term averages for this period of time. did indicate the necessity for station diversity, a matter that was examined in the northern experiment [Goodman et al., 1997]. While Jull found discrepancies between model predictions and experimental data during disturbances, he did note that models are most useful during undisturbed periods. Even under these conditions it is important to find ways to extrapolate ionospheric data successfully. This is because we are generally dealing with an undersampled environment. 6. Observations [17] The Svalbard to Barrow path is an interesting one since it traverses the polar cap. Figure 1 gives the geometry. Figure 2 is a sample model output. Figures 3 and 4 are examples of normal and abnormal HF propagation across the pole that was obtained in December Just how abnormal is the abnormal condition? Figures 5 10 contain estimates of the MOF for all runs for the data period between 1 6 December 2000, 6 28 February 2001, July 2001, August 2001, November 2001, and December This total sample is 112 days, and it is composed of 64,000 ionograms. [18] Figure 2 is an output from the Voice of America HF propagation code called VOACAP [Teters et al., 1983] (see also for the December period and high sunspot number. We see that the maximum usable frequency (MUF) is rather constant, being the order of MHz all day long. This is a median value. The method used is one that displays the number of MUF days as well as the MUF itself. This is the percentage of time that a predicted maximum observed frequency (i.e., predicted MOF) exceeds the MUF for any specified hour during the month. On average, we see that a predicted MOF of 22 MHz should be observed about 20% of the time, or for 6 MUF days reckoned over a month. By inspection, the upper decile is approximately 24 MHz. Shortly we shall see that a change in some of our assumptions can lead to an increase in the MUF and the upper decile values. [19] The reader should recognize that the so-called model MUF at a given hour is a median for the given Figure 8. Autoscaled estimates of the MOF for August The diurnal patterns are consistent with long-term trends for the first 40% of this set of data. For the last 50% of the data set, there is a distinct 5 day oscillation in the MOF pattern. 7of12

8 Figure 9. return. Autoscaled estimates of the MOF for November The large diurnal variations month and is derived from VOACAP and dependent upon the ionospheric coefficient set used. The observed MOFs at any given hour can be compared with the monthly MUF, but one would not expect individual daily values to be in agreement except by chance. Naturally, the aggregate of all observed daily values might used to construct a median that could be more reasonably compared with the model MUF value. [20] The EMOF events discussed in this paper should not be confused with low-level above-the-muf scatter signals that may always be observed if system sensitivity is sufficiently robust. The present observations, including the EFoP events, are associated with reflection of HF radio waves. Above-the-MUF (ABM) scatter propagation has been documented by Wheeler [1966], but Phillips and Abel base the theory upon earlier unpublished work during Project EARMUF undertaken for the U.S. Army Signal Corps in 1958 (G. Lane, private communication, 2003). The theory postulates a Gaussian distribution of patches and blobs that enable ABM propagation to occur. The original theory has been represented as a quasi-empirical model based upon midlatitude observations. It is noteworthy that the ABM model is not incorporated in the Lucas-Haydon tables cited in this paper (i.e., Tables 1 and 2), but it can be invoked from VOACAP if desired. It turns out that the median of all MOFs observed for the December period is 17.5 MHz and that this is virtually the same as VOACAP predicts as the MUF for the same period and conditions. [21] As an illustrative example, we have developed the MOF distribution for the period in December 2000 (i.e., Figure 5). We were especially interested in examination of the distribution of EMOF events. An automatic Figure 10. Autoscaled estimates of the MOF for December This is a 5 day period. Compare this with the 6 day period in December There are some similarities but the diurnal variations are not as intense. 8of12

9 Table 1. Distribution of F2 Layer MUFs at 80 North Local Time Winter/high Equinox/high Summer/high scaling algorithm selected the MOF for each ionogram. We cataloged all MOFs, including normal values representative of climatology as well as EMOF events, without distinction. The MOFs of all categories were characterized by high signal strength echoes. Typically, they exhibited distinct nose-like features consistent with a classical junction frequency (composed of the merger of high and low rays), although on some occasions the high ray would not be observed. Figure 11 contains the MOF distribution for 1 6 December. The distribution is the number of samples for which the MOF index, a convenient parameter used in the figure, is exceeded. There are 1280 valid samples. It should be noted that there are 170 samples exceeding an index of 15 (or F = 28 MHz), and the upper decile is characterized by 128 samples. We estimate the upper decile to be 30 MHz, but it could have been higher had the instrumentation not been limited to readings of 30 MHz. The observed median frequency is 17.5 MHz (an index of 9.8), the same as the predicted MUF from VOACAP. On the basis of an empirical model from Lucas and Haydon [1966], and discussed below, the expected upper decile is approximately 1.38 times the median (or MUF) under similar conditions. This corresponds to a Lucas-Haydon value of 24.5 MHz for the predicted upper decile. The observed value is 5.5 MHz higher. We maintain that the EMOF events explain the difference, and that polar patches constitute the mechanism. [22] It is of interest to examine the MUF distributions as determined by Lucas and Haydon [1966]. The distribution of the daily values of the F2 layer MUF about their monthly medians is given in tables of ratios of the decile values to the median value. These values are a replacement for the often-used values of 0.85 (for the optimum working frequency (FOT)) and 1.15 (upper decile). The full tables were developed as a function of season, sunspot number, local time, and every 10 degrees of geographic latitude. The highest latitude given is 80 degrees north, and this is in the neighborhood of the ray trajectory midpoint (namely Svalbard to Barrow). Table 1 is a listing of the lower-upper decile range for specified conditions extracted from Lucas and Haydon [1966]. [23] We see that the variation from the MUF is generally least in the summer and greatest in the equinox and wintertime. From Table 2 we find that the decile values of the MUF distribution are not too different from the classical values (i.e., ) for midlatitudes. [24] Lucas and Haydon [1966] had reason to believe that the highest frequencies of operation would be elevated in the polar cap in comparison to subauroral MUF distributions. Does their empirical representation arise due to the presence of blobs and patches of ionization? One can only guess that this is the case. There is some diurnal variation in the 80 N data shown in Table 1. For the Svalbard-to-Barrow path we assume that this variation would be washed out, given the fact that the midpath of the ray trajectory is near the geographic pole. For this case the average decile values are computed to be 0.70 and 1.38 as a fraction of the median MUF. If the MUF were taken to be 17.5 MHz for December, then the upper decile would be MHz. Lane [2001], the leading expert on IONCAP and VOACAP models, has indicated that use of International Radio Consultative Committee (CCIR) coefficients should be preferred, and this utilization yields a MUF of 17.8 MHz and an upper decile of 24.2 MHz. This is very close to our assessment. Lane also indicates that sporadic E contributions will tend to raise the MUF somewhat. An equivalent oblique-incidence ionogram can be generated from VOACAP method 25, and using CCIR coefficients and sporadic E contributions, as suggested by Lane, Table 2. Distribution of F2 Layer MUFs at 40 North Local Time Winter/high Equinox/high Summer/high of12

10 Figure 11. MOF distribution for 1 6 December The median frequency is seen to be 17.5 MHz (index = 9.8). The Lucas-Haydon model for similar conditions gives the upper decile as 1.38 times the median. This corresponds to a Lucas-Haydon value of 24.5 MHz for the upper decile. The upper decile read from the figure is in excess of 28 MHz (i.e., index >15). we obtain a MUF of 20 MHz at 1100 UTC. Using the revised figures, the predicted upper decile would be 1.38 (20 MHz) = 27.6 MHz. [25] We may conclude the following about the database used by Lucas and Haydon: (1) either it dos not include the features responsible for the EMOF events we describe in this paper, or (2) the authors smoothed out these features in the development of their MUF representations. Eventually this incomplete picture, vis-à-vis polar patches, found its way into all current models that exploit the two established sets of ionospheric coefficients (i.e., CCIR and International Union of Radio Science (URSI)). It is recommended that future models account for these new features, albeit statistically. Real time models using sensor updates may be able to account for the features dynamically. 7. General Discussion [26] The existence of elevated MOF values represents an increased opportunity for communication across the pole. The additional propagation windows or modes allow for the use of bands that are traditionally unexploited. However, there are a number of issues raised by these new propagation windows. They include: (1) communication signal quality using these modes, (2) temporal persistence of the modes, (3) spatial dimension of the modes, and (4) predictability of the modes. [27] The issue of how well analog voice and digital transmissions will perform using the polar EMOF channel (i.e., patches or blobs) will depend upon the nature of 10 of 12 the ionospheric scattering regions. We have observed rather coherent signals as well as fuzzy patterns resembling spread F echoes. The spread returns may be problematic, and one would expect rapid fading to be introduced. These signals should be studied to ascertain the properties such as Doppler spread and multipath spread. Both narrowband (i.e., <12 khz and wideband (i.e., >12 khz) channel parameters should be examined. An initial study should involve the transmission of voice and data over standard 3 khz channels to determine articulation scores (for voice) and the bit error rate (for data). The oblique-incidence chirp sounder data should be augmented by vertical incidence data in close proximity to the scattering center, if possible. Sounder data of opportunity may serve as a proxy, and Thule is a possibility. Other corroborative data might include SuperDARN data sets. [28] Continued data collection will lead to more information concerning issues 2 4 above. The current data set will provide a first step in this regard. A network of chirp sounder systems will provide for improved spatial sampling. We certainly need several simultaneously monitored paths across the pole to provide a better picture of the polar EMOF structures. Otherwise, we would have to rely on other data. We have plans to instrument a path between Reykjavik (Iceland) and Barrow and between Reykjavik and Svalbard. [29] Predictability is a major issue. With a good forecasting capability it would be possible to plan for use of the polar EMOF structures when they are thought to have an increased likelihood of occurrence. This capability, coupled with a nowcasting capability based

11 upon direct measurement of the EMOF patterns, will enable a dynamic frequency management system to be developed. This is our ultimate goal. A good system could be developed based upon real time ionospheric measurements alone, using basic understanding of the temporal evolution of the structures and trend line analysis. However, the solar-terrestrial environment is changing all the time, and evolving sources for emergent polar EMOF patterns need to be folded in all the time. In short, we need to be cognizant of new patterns that penetrate the space-time domains for the HF communication paths of interest. 8. Practical Consequences of Polar EMOF Structures [30] A major consequence of the polar EMOF structures includes the potential for use of much higher bands than are currently in use in the polar region. This has been discussed in the introductory portion of this paper. The use of higher bands, if available, will reduce the atmospheric noise competing with signal and will reduce the amount of ionospheric absorption inflicted on the signal. To accomplish this, regulators must work with radio engineers to allow more spectrum use in the polar regions. In some cases, HF bands in the range between 18 and 30 MHz are authorized for use for a given service (e.g., maritime-mobile or aeronauticalmobile) but are not licensed since the requirements are derived on the basis of climatology. In addition, of course, polar blobs (and polar EMOF signals) are not a part of the archived climatology. Hence the studies described in this paper are important as a means to educate system engineers who must field systems in the polar environment. [31] Acknowledgments. The authors would like to acknowledge Ed Goldberg for software support and maintenance of the data base. John McMains set up the transmission facilities in Isfjord, Svalbard, and he coordinated the RPSI activity with Vivianne Jodalen of Norwegian Defense Research Establishment and Per Krokan of Telenor, Svalbard. We would also like to thank the Norwegian Post and Telecommunication Authority for providing a license to transmit during the test period. References Benson, R., and J. M. Grebowski (1999), Extremely low ionospheric peak altitudes Possible relationship to polar holes, in Proceedings of the 1999 Ionospheric Effects Symposium, edited by J. M. Goodman et al., pp , Off. of Nav. Res., Arlington, Va. (available as PB from Bishop, G. J., J. A. Klobuchar, A. E. Ronn, and M. G. Bedard (1989), A modern trans-ionospheric propagation sensing system, in Operational Decision Aids for Exploiting or Mitigating Electromagnetic Propagation Effects, NATO- AGARD-CP-453, AGARD, San Francisco, Calif. Buchau,J.,B.W.Reinisch,E.J.Weber,andJ.G.Moore (1983), Structures and dynamics of the winter polar cap F region, Radio Sci., 18, Crowley, G. (1996), Critical review of ionospheric patches and blobs, in Review of Radio Science , pp , Oxford Univ. Press, New York. Davies, K. (1990), Ionospheric Radio, IEE Electromagn. Wave Ser., vol. 31, Peter Peregrinus Ltd., Inst. of Electr. Eng., London. Goodman, J. M. (1991), HF Communications: Science and Technology, Van Nostrand Reinhold, New York. Goodman, J. M., and J. W. Ballard (1999a) Dynacast-assisted frequency management for HF communication and broadcasting systems, in Proceedings of 1999 Ionospheric Effects Symposium, edited by J. M. Goodman pp , Off. of Nav. Res., Arlington, Va. (available as PB from Goodman, J. M., and J. W. Ballard (1999b), Method and apparatus for real time ionospheric mapping and ionospheric forecasting, Patent , U.S. Patent and Trademark Off., Washington, D. C. Goodman, J. M., J. Ballard, and E. Sharp (1997), A long-term investigation of the HF communication channel over middle-and high-latitude paths, Radio Sci., 32(4), Hunsucker, R. D. (1983), Anomalous propagation behavior of radio signals at high latitudes, in Propagation Aspects of Frequency Sharing, Interference and System Diversity, AGARD CP 332, vol. 11, edited by H. Soicher pp , Advis. Group for Aerosp. Res. and Devel., Paris. (available as ADA from Natl. Tech. Inf. Serv., Springfield, Va.) Jull, G. W. (1964), HF propagation in the Arctic, in Arctic Communications, edited by B. Landmark, pp , Pergamon, New York. Kelley, M. C. (1989), The Earth s Ionosphere: The Plasma Physics and Electrodynamics, Academic, San Diego, Calif. Landmark, B. (Ed.) (1964), Arctic Communications, Pergamon, New York. Lane, G. (2001), Signal-to-Noise Predictions Using VOACAP, Rockwell Collins, Cedar Rapids, Iowa. Leid, F. (1967), High Frequency Radio Communications with Emphasis on Polar problems, AGARDograph 104, Technivision, Maidenhead, UK. Lucas, D. L., and G. W. Haydon (1966), Predicting statistical performance indices for high frequency telecommunication systems, ESSA Tech. Rep. IER 1-ITSA-1, U.S. Dept. of Commerce, Boulder, Colo. Rodger, A. S. (1998), Polar patches Outstanding issues, in Polar Cap Boundary Phenomena, NATO ASO Ser., vol. 509, edited by J. Moen, A. Egeland, and M. Lockwood, pp , Kluwer Acad., New York. 11 of 12

12 Rodger, A. S. (1999), Recent advances in geospace research, in Review of Radio Science , edited by R. Stone, pp , Oxford Univ. Press, New York. Teters, L. R., J. L. Lloyd, G. W. Haydon, and D. L. Lucas (1983), Estimating the performance of telecommunication systems using the ionospheric transmission channel, IONCAP user s manual, NTIA Rep , U.S. Dept. of Commerce, Washington, D. C. (available as PB from Natl. Tech. Inf. Serv., Springfield, Va.) Wheeler, J. L. (1966), Transmission loss for ionospheric propagation above the standard MUF, Radio Sci., 11, J. W. Ballard, Radio Propagation Services, Inc., 199 First Street, Suite 326, Los Altos, CA 94022, USA. J. M. Goodman, Radio Propagation Services, Inc., 8310 Lilac Lane, Alexandria, VA , USA. (jm_good@ cox.net) 12 of 12