First results of mapping sporadic E with a passive observing network

Transcription

1 SPACE WEATHER, VOL. 9,, doi: /2011sw000678, 2011 First results of mapping sporadic E with a passive observing network D. D. Rice, 1 J. J. Sojka, 1 J. V. Eccles, 1 J. W. Raitt, 1 J. J. Brady, 1 and R. D. Hunsucker 2 Received 15 March 2011; revised 22 August 2011; accepted 28 September 2011; published 14 December [1] Sporadic E (E s ) can have dramatic effects on communications in the HF and low VHF range, producing over-the-horizon propagation for signals normally restricted to line-of-sight, and sometimes blocking F region propagation of signals in the lower HF range. Measuring the E region winds believed to produce E s is difficult, and no practical means of predicting E s occurrence currently exists other than statistical models. We describe a low-cost observing network based on software-controlled receivers that continuously watches for E s in near-real time using oblique HF propagation from existing transmitters. Results from an 11-day pilot campaign in July 2008 demonstrated that even a limited number of receivers in the network can readily determine the presence and extent of E s patches. These observations indicate that E s often develops quickly over regions of several hundred kilometers rather than gradually drifting across an area. These widespread E s blooms have been observed near winter solstice and occasionally at other times of the year; their lifetime depends on the season but can be several hours during the summer. The current network allows the extent of E s in portions of North America to be evaluated: the geographical distribution of E s and bounds on the density of the layer are inferred from its effects on the ionospheric maximum usable frequency (MUF). This study demonstrates quantitatively that E s mapping can provide information about E s layer geographical growth and decay. The observed sudden widespread E s blooms are space weather events that can have significant impact on HF/lower VHF communications and propagation model predictions. Citation: Rice, D. D., J. J. Sojka, J. V. Eccles, J. W. Raitt, J. J. Brady, and R. D. Hunsucker (2011), First results of mapping sporadic E with a passive observing network, Space Weather, 9,, doi: /2011sw Introduction [2] Sporadic E (E s ) and its effects has been observed for decades. Dense E s layers typically occur in the same height range as normal daytime E, km, but with much greater critical frequencies than daytime E layers: while a daytime E layer critical frequency might be on the order of f o E 4 MHz, midlatitude E s often produces f o E s > 10 MHz, and may occur day or night. E s is also observed forming at higher altitudes ( 140 km) and gradually descending to normal E heights, where it may develop become more dense. [3] The greatest impact of E s is on HF (3 30 MHz) and lower VHF ( MHz) communications. For lower HF (<10 MHz), the impact is generally seen at night, when signals at these frequencies typically travel thousands of kilometers due to F region reflection. When E s appears, it can block the F region path at either end, causing degradation or loss of the signal. At higher frequencies, E s typically causes unexpected long-distance propagation, which can have both positive and negative consequences. On the 1 Space Environment Corporation, Providence, Utah, USA. 2 RP Consultants, Klamath Falls, Oregon, USA. positive side, radio amateurs exploit E s openings to communicate over thousands of kilometers at frequencies that are normally limited to line-of-sight distances [Neubeck, 1996]. On the negative side, broadcasters and point-to-point communicators who expect to reach local listeners may find themselves experiencing interference from distant transmitters during E s openings. [4] In all of these cases, E s causes signal propagation to behave very differently from what is expected from routine experience or computer propagation models. E s has the greatest impact on predicted propagation during the midlatitude summer E s season, and at high latitudes where auroral sporadic E is common. These anomalies can have serious consequences for communicators participating in disaster response, long-haul aviation (especially transpolar flights), HF geolocation, and other services where reliable HF/VHF propagation is crucial. [5] Few computer models attempt to account for E s because of the complexity of the phenomenon. It is generally believed that midlatitude E s is due to wind shears [Whitehead, 1989], modulated by gravity waves [Yokoyama et al., 2004], tides, and planetary waves [Schunk and Nagy, 2000; Haldoupis et al., 2004]. The winds in the E region are Copyright 2011 by the American Geophysical Union 1of11

2 difficult to measure and to model, so at present E s propagation cannot be modeled or predicted with any certainty. Instead, E s is typically described in terms of seasonal and diurnal statistical trends, e.g., Smith [1976], if it is addressed at all. Thus the question remains: how can E s propagation be specified in a way that is practical and useful? [6] Traditionally, E s has been studied with active instruments such as ionosondes and incoherent scatter radars [Mathews, 1998]. These instruments can provide a detailed height profile for the ionosphere above the instrument, with time resolution of a few minutes. Such studies have provided much of what is known about E s ; for example, the global distribution of E s was first inferred from ionosonde network observations [Smith, 1962]. Unfortunately, the modern network of sounders and radars is quite sparse compared to the IGY network (see, e.g., uml.edu/stationmap.html), making detailed geographical studies of E s behavior with these instruments impractical. [7] More recently, GPS occultation studies have been performed by low-earth-orbit (LEO) satellites that allow the presence of E s to be inferred over wide areas set by the satellite orbital geometry with respect to the GPS constellation [Wu et al., 2005; Garcia-Fernandez and Tsuda, 2006]. This method allows global E s distributions to be mapped, but a given area may not be revisited for hours, depending on the satellite orbit. Also, since the geometry of the satellites is constantly changing, it is difficult to specify how the extent of an E s region might be changing over time. [8] The method for mapping E s described here provides a means to study the mesoscale morphology of E s with good time resolution [Eccles et al., 2008]. HF beacon transmitters are monitored by a network of inexpensive receivers, and paths that have propagation well above the expected maximum usable frequency (MUF) are assumed to be due to E s. This assumption has been tested against ionosonde observations on one of the monitored paths, and good agreement has been found between ionosonde E s measurements and inferred E s propagation. Propagation data has been collected beginning in 2002 and continuing through the present, including the large variations in the previous solar cycle and the extremely mild start of the new cycle. One major advantage of the HF monitoring method is its low cost, with individual monitors costing about U.S. $5000 [Sojka et al., 2005]. Amateur monitoring results may also be included in the analysis. [9] The development of E s has been observed during summer E s seasons using this monitoring technique. E s is seen to cover regions of longitude in about an hour. Regions may expand east and west; regions often develop in the eastern United States during local morning hours, and then are seen in the western United States a few hours later. However, regions also develop independently in the eastern and western United States. [10] Mapping the occurrence of E s using a widely distributed array of inexpensive instruments has many practical applications for improving space weather specifications and notifications to both amateur and professional communicators. Propagation models could be modified to ingest current E s conditions, or at least warn users that current predictions might be unreliable. HF communicators can be made aware of the presence of E s that might affect the quality or availability of communications facilitated by automatic link establishment (ALE) protocols. At VHF frequencies, E s mapping would allow potential interference to local stations from distant FM and television broadcasters to be identified [Miya, 1996]; historically, analog television signals arriving via E s could be identified fairly readily, but with modern digital television, such interference causes degradation or loss of the local signal, and might be attributed to any number of other causes. Es maps would provide clues for diagnosing intermittent television service failures. 2. Sporadic E Propagation [11] Space Environment Corporation (SEC) has deployed a network of Space Weather-Aware Receiver Elements (SWAREs) to monitor radio signals from VLF through HF [Rice et al., 2009] following earlier HF monitoring efforts [Eccles et al., 2005]. The SWAREs produce their own threedimensional ionospheric models based on International Reference Ionosphere (IRI) F region [Bilitza, 2001], combined with the DDDR (Data Driven D-Region) [Eccles et al., 2005] model for the E and D region. Current geophysical indices are obtained via the Internet. The HASEL raytracing program [Coleman, 1993] is used with the model ionosphere to estimate the expected signal strength from the standard time station WWV (near Fort Collins, Colorado) between 2.5 and 20 MHz at the receiving site and estimate signal absorption. [12] The primary means of inferring the presence of E s with this system is the observation of HF propagation at frequencies well above the MUF expected according to the DDDR/IRI ionospheric models. Signal observation entails identifying the distinctive tone patterns of WWV, or decoding the digital identification of amateur transmissions. As the number of monitored paths increases, the use of detailed ray tracing through the model ionosphere becomes a rather cumbersome method for determining the MUF in order to distinguish normal E/F propagation from potential E s propagation. A simpler methodology was developed to test for E s on a given path: the midpoint of the single-hop path from the transmitter to the receiver is determined, and the maximum reflection frequency is determined for the path using the secant law [Hunsucker, 1991] at each height in the model ionosphere up to the F region peak. The overall maximum reflection frequency is then taken to be the MUF. This approach neglects refraction by underlying ionization for F region paths, and generally overestimates the MUF compared to ray-tracing MUF estimates for the same model ionosphere. 2of11

3 Figure 1. Maximum usable frequency computed by the geometrical secant law (Geo) and ray tracing (RTr) for the paths from WWV in Fort Collins, Colorado, to receivers at Bear Lake Observatory (BLO), Utah, and Klamath Falls, Oregon (KFO). Calculations were based on July 2008 conditions. [13] Figure 1 shows the comparison between the geometrical secant law calculation (Geo) and the ray-tracing analysis (RTr) for two WWV paths in July The path to Bear Lake Observatory (BLO) is shorter (550 km) than the path to Klamath Falls, Oregon (KFO), which is 1400 km. The simpler geometric calculations are consistently at least 0.5 MHz higher than the ray-traced estimate in the early morning, and at least 1 MHz higher than the ray-traced estimate in the afternoon. Thus the simplified calculation provides a conservative limit for judging signals to be above MUF. For the conditions shown here, it is reasonable to say that no propagation of the 20 MHz WWV signal is expected at either receiver site, and 15 MHz signal is not expected at BLO. If such signals are detected, then they are assumed to be due to E s. This assumption is reasonable in the absence of significant positive-phase geomagnetic storms, which might raise the MUF above model estimates. This assumption is discussed further in the next section. [14] The WWV-KFO path geometry is such that the path midpoint occurs near BLO in Utah, where a low-power Canadian Advanced Digital Ionosonde (CADI) is operated by SEC. The CADI allows E s layers to be quantified and compared to the layers inferred from the HF signal monitoring. The ionosonde also allows any geomagnetic storm effects to be observed and quantified; however, there has been little storm activity during the prolonged solar minimum in which this study has been performed. 3. Path Geometry and Sporadic E Sensitivity [15] For a particular E s reflection height he s, a given E s layer density will support propagation for shorter paths as the signal frequency decreases. Figure 2 illustrates the minimum single-hop transmitter-to-receiver path length supported for different combinations of E s critical frequency and signal frequency over the range of reasonable E s heights. This minimum path length is known as the skip zone radius, because receivers closer than this distance to the transmitter will be skipped by rays reflected from the specified E s layer. For each signal frequency, the upper curve in the shaded region corresponds to he s =140km,and the lower region boundary corresponds to he s = 100 km. [16] For example, a modest E s critical frequency ( f o E s )of 5 MHz and he s = 100 km will propagate 10 MHz signals on paths greater than 350 km, while 20 MHz signals will only propagate on paths greater than 900 km, and 30 MHz signals will not propagate (the maximum single-hop path length being limited by the curvature of the earth.) Doubling f o E s to 10 MHz, not uncommon during midlatitude summer E s season, supports 30 MHz signal propagation for paths greater than 600 km. 3of11

4 Figure 2. Minimum transmitter-to-receiver path length supported for various signal frequencies and E s critical frequencies. The layer E s height varies from 100 km (bottom of shaded region) to 140 km (top of shaded region). [17] Figure 2 also indicates the density/height ambiguity inherent in this simple technique: if the skip zone radius of a 30 MHz signal was found to decrease from 875 km to 600 km, it could be interpreted as a sporadic layer with f o E s = 10 MHz descending from 140 km to 100 km height, or as a layer with he s = 100 km increasing in density from f o E s = 7.7 MHz to 10 MHz, or as some combination of density and height variations within this range. Observations of other transmitter/receiver paths in the same region may be used to reduce this ambiguity. [18] If an observed signal frequency is below the estimated MUF for the path, then the means of propagation (E s, E, F) cannot be determined without additional information. Also, for frequencies below about 15 MHz, daytime D-region absorption becomes a factor; E s paths have low elevation angles and subject the signals to long D-region transits, which may absorb much of the signal energy, and such signals may actually fade as the result of E s formation. Absorption effects on monitored signals have been studied in more detail in Eccles et al. [2005, 2008]. Thus there is a tradeoff in selecting frequencies to be monitored for E s propagation such that they will be high enough to avoid absorption and be above the expected MUF for the path, but low enough in frequency to propagate at moderate E s densities. These criteria depend on solar flux, but in general frequencies below 15 MHz are avoided for E s analysis due to the added complications of daytime D-region absorption at the lower frequencies. [19] Manual analysis of data sets has shown cases where signals at frequencies below the MUF are lost due to E s. This loss may be due to blocking the normal F path at the transmitter or receiver side. Daytime E s may cause the rays to follow low-angle paths which experience high D-region absorption, causing deep fading or signal loss. However, there are too many factors which may cause unexpected loss of signal, such as severe interference or strong X-ray flares. In one case, prolonged signal loss at BLO was found to be due to a rodent chewing through the antenna cable near the mast. Thus we do not consider the loss of below- MUF signals in the automatic E s mapping procedures. [20] The prolonged solar minimum conditions that lasted from 2006 into 2011 were excellent for studying E s via HF propagation analysis because the MUFs were relatively low and stable (few positive-phase ionospheric storms occurred to raise the MUF beyond model estimates.) E s could be inferred with confidence at frequencies down to 15 MHz on some paths. However, as solar activity increases and MUFs rise, E s observation will be limited to shorter paths and higher frequencies, and MUFs will have to be checked carefully against ionosonde observations for storm disturbances. During the more active conditions prior to 2006, E s determination was generally restricted to frequencies of 20 MHz and higher, depending on path length. [21] During more active conditions, a conservative approach would be to discard observations during and after significant storms. A storm-time increase in the F-layer peak density beyond model expectations can be 4of11

5 Figure 3. SWARE sites and WWV path midpoints for July 2008 campaign. WWV is located near Fort Collins, Colorado, and transmits at 2.5, 5, 10, 15, and 20 MHz. When E s propagation is detected, the layer is assumed to be at the single-hop path midpoint. confirmed by either ionosonde or GPS TEC measurements. An ionosonde will reveal local f o F 2 increases, while the regional GPS TEC observations will show enhancements from which f o F 2 changes can be inferred. [22] Storm-related F-layer enhancements can be quite large; the dayside F region height increases and in sunlight its density can be increased, both factors affecting the MUF. In extreme cases, the TEC can be doubled, corresponding to nearly doubling N m F 2 and an increase of about 40% in f o F 2. These positive phases only occur at the beginning of storms and are only seen in dayside longitudes [David and Sojka, 2010]. When evidence of such changes is observed in ionosonde or TEC data, the E s interpretation should be discontinued until observations are again consistent with model values. 4. Western United States Sporadic E Campaign [23] A campaign was undertaken in July 2008 to observe E s propagation across the western United States using SWARE HF signal measurements, amateur radio data, and CADI ionograms. The SWAREs were operated at various locations for periods ranging from a day to more than a week. Locations are shown in Figure 3 and summarized in Table 1. The minimum f o E s specified in Table 1 assumes 20 MHz reflection at 100 km; higher E s layers would require somewhat higher E s critical frequencies to support 20 MHz signal propagation from WWV, as indicated in Figure 2. [24] The 20 MHz WWV signal received at KFO between 4 July 2008 (day 186) and 14 July (day 196) is compared to E s measured near the path midpoint in Figure 4. The E s critical frequency f o E s measured by the CADI at BLO is shown as the red trace, indicating modest summer E s densities. The WWV signal strength is shown as the colorcoded band; dark areas represent strong signals and white represents a signal-to-noise ratio of zero. The band s position at 3.88 MHz represents the minimum f o E s required to reflect 20 MHz from WWV to KFO in a single hop for hes = 100 km. A good correlation is seen between Table 1. SWARE Locations, July 2008 Campaign SWARE Site Location Latitude ( N) Longitude ( E) Minimum f o E s (MHz) WWV Path (km) ABY Albany, OR BKR Baker, OR BLO Bear Lake Obs., UT ERK Eureka, NV HWT Hawthorne, NV KFO Klamath Falls, OR PLK Priest Lake, ID PRV Providence, UT PVL Prineville, OR RNO Reno, NV RWY Rockaway, OR TUC Tucson, AZ WEN Wendover, NV of11

6 Figure 4. WWV 20 MHz signal strength (dark is maximum strength) compared to f o E s measured at BLO (red trace.) Signal strength is positioned at 3.88 MHz, the minimum f o E s required to reflect the 20 MHz signal between WWV and KFO. strong WWV-KFO 20 MHz signals (dark regions in the band) and CADI f o E s measurements exceeding the 3.88 MHz single-hop E s path threshold. This correlation provides evidence supporting our assumption that E s is responsible for propagation at frequencies above the estimated MUF. [25] In general, signal propagation is observed at KFO when f o E s is near or above the calculated threshold. The BLO CADI is located about 160 km east of the WWV- KFO path midpoint, so the agreement indicates that the observed E s regions cover at least 160 km longitudinally, and also that the regions appeared at nearly the same time over BLO and at the WWV-KFO midpoint. [26] On day 188, f o E s measured by the BLO CADI never reached 3.88 MHz, hence no propagation is expected for the WWV-KFO 20 MHz path. Indeed, day 188 was the least active day of the campaign, as seen in Figure 4. Toward the end of day 188, E s activity over BLO began increasing, and by day 190 f o E s exceeded 3.88 MHz for most of the day; the WWV-KFO 20 MHz signal is correspondingly strong through that day. The observed signal strengths for paths that require E s for propagation are shown in Figure 5 for the quiet day 188, and in Figure 6 for the more active day 190. [27] Day 188 shows no large-scale activity over the area, with only isolated instances of E s during the day and with the suggestion of increasing activity toward the end of the day. The lack of consistent propagation demonstrates that the normal E/F MUF is indeed lower than the 15 and 20 MHz signal frequencies on these paths. [28] Just two days later, propagation is completely different (Figure 6) and signal is even detected on the WWV- PRV 20 MHz path, which requires an f o E s of at least 6.91 MHz. One interesting feature that is seen in much of the data is that E s propagation ceases around dawn (about 11:30 UT) and resumes mid-morning (around 15:00 UT in this figure), suggesting that sunrise dynamics disrupt the relatively stable E s layer for several hours. In terms of path midpoints (Figure 3), E s covering Utah from the Idaho border down to the Four Corners region of New Mexico would provide the observed propagation early in day 190. The inclusion of ABY, BKR, and PLK around 03:00 UT could be explained by expansion of the E s region northward into Idaho and Wyoming. 6of11

7 Figure 5. WWV signal strength (dark is maximum strength) received on day 188 of The receiver site (see Table 1) and frequency is given on the right margin for each signal band. [29] E s is present for the east (PRV) and south (TUC) receivers at the start of day 190. It is also present less consistently at the southern receivers in the west (RNO and KFO), then is seen somewhat north and west (ABY), and finally much further north (BKR and PLK) by about 03:00 UT where it remains in place until dawn. The resumption of E s propagation occurs across the region around 15:00 UT. (Note that RNO is shut down at 15:30 UT and ABY is shut down at 17:00 UT for relocation.) Curiously, the southernmost (TUC) and northernmost (PLK) receivers see propagation resume almost simultaneously, together with less consistent signals at western receivers ABY and KFO. RNO appears shortly after, then PRV, and finally BKR. The map suggests separate regions over Wyoming, Idaho, and New Mexico appeared almost simultaneously shortly before 15:00 UT and merged over northern Utah/southern Wyoming shortly after 15:00 UT. [30] In describing the development of E s and its geographic evolution, it is necessary to consider whether the E s region is expanding, drifting, or if the density of the region is changing. This interpretation may be difficult when observations are made at a single fixed frequency. If a stationary E s patch exists with f o E s slightly below the value required to support propagation of the fixed frequency path, a small increase in density (and increase in f o E s ) would cause the patch to suddenly appear on the propagation map. However, if the map contains many different signal frequencies and path lengths, density changes may be deduced. [31] Animated maps of the data represented in Figures 5 and 6 have been prepared which indicate the presence or absence of E s at the path midpoints shown in Figure 3; these maps are available through the SEC website. The geographic coverage of this early campaign is limited to central Utah through southeastern Idaho, but could be increased by adding additional receivers and monitoring other HF beacon transmitters. This expansion has been pursued and is discussed in the next section. [32] A careful analysis of the July 2008 observations showed only one instance of apparent motion of the E s region. On 4 July 2008, E s propagation gradually extended 7of11

8 Figure 6. WWV signal strength (dark is maximum strength) received on day 190 of westward through the local afternoon, with an approximate speed of 70 m/s. This velocity is consistent with other studies; Whitehead [1989] stated that most reported E s motions are in the m/s range. However, the other E s observations showed propagation commencing approximately simultaneously throughout the observing area. 5. Sporadic E Morphology Across the United States [33] Additional observations of amateur radio 28 and 50 MHz transmissions have been added to the analysis to provide a picture of the development of E s across the United States. These observations are collected in near real time from the PropNET project [Ford, 2008], which maintains a database of amateur radio digital transmissions received at numerous sites around the world. The database includes the frequency and coordinates of the receiver and transmitter. This information is filtered using the MUF estimate described above, and paths are limited to those consistent with single-hop E s propagation in the contiguous United States (CONUS). [34] In order to show the development of E s over time, the number of probable E s paths is counted in bins covering 1 h UT and 5 longitude across the CONUS latitudes. It should be noted that the lack in observations in the Midwest, E s is largely due to the reduced number of suitable paths (fewer transmitters/receivers) in the Midwest compared to the more populous east and west coasts. Most stations participating in PropNET are automated and operate continuously, so temporal behavior does not reflect the schedules of human operators. Examples of the resulting time history plots are shown in Figure 7. [35] On 4 May 2010 (Figure 7, left), E s is present at 0 UT in the west (late afternoon) and east (early evening.) The western E s fades during the evening (5 UT) but returns the next morning, 11 UT, and persists until local noon, 20 UT. The eastern E s contracts to a small region in the southeast that persists through the night and fades during the morning. The latitudinal behavior may be found in more detailed maps that are generated automatically ( Two examples of these maps are shown in Figure 8, with E s spread across 8of11

9 little E s is found in the west until the next morning, around 14 UT. In this case, the E s appears to spread eastward, with east coast E s peaking in the early afternoon as west coast E s is fading. [37] There is a general tendency for E s to intensify in an east-to-west direction, consistent with semi-diurnal tides [Arras et al., 2009; Christakis et al., 2009]. However, as seen in Figure 7 (right), a west-to-east growth is sometimes observed. Often, a strong E s region is observed only in the east or only in the west on a given day. In latitude, E s is observed most often in the southeast and southwest, and appears to spread north as E s intensifies in the southern regions. [38] In order to view trends over longer periods, we sum the bins in Figure 7 over a range of longitudes and display the hourly total count for the region over days of The results are shown in Figure 9, with one region covering the western United States (232.5 E E (left)) and another region covering the eastern United States (272.5 E E (right)). The central region generally has fewer suitable paths, as discussed above with respect to Figure 7, so it is omitted from the summary plots. [39] In the west, the occurrence of probable E s propagation peaks around 18:00 UT with a weak peak around 02:00 UT. In the east, the peaks are around 15:00 UT and 23:00 UT. These times correspond to midmorning and Figure 7. E s propagation at 28 and 50 MHz for 5 bins covering the contiguous United States on (top) 4 May 2010 and (bottom) 10 May much of the United States in the evening (01:40 UT, top) but restricted to the southwest the following morning (13:40 UT, bottom.) [36] On 10 May (Figure 7, right), a similar persistent E s region is found in the southeast during the evening, but Figure 8. Maps of E s from 4 May (top) Map for 01:40 UT shows E s scattered across the United States, corresponding to Figure 7, bottom left. The curved line is the evening terminator. (bottom) Map for 13:40 UT shows E s primarily in the southwest, with the morning terminator moving into the Pacific. 9of11

10 Figure 9. Occurrence of probable E s propagation in the western United States (left) and eastern United States (right) during the summer of early evening in each region. E s vanishes on most nights and reappears after local sunrise. Including the full range of U.S. latitudes in the region binning undoubtedly smears out any sunrise/sunset effects. 6. Discussion [40] Detailed mapping of E s across a wide geographical region is an ongoing challenge that has not yet been fully addressed. Successful mapping of E s would enable significant scientific progress to be made in understanding the formation and dynamics of the phenomenon and would potentially put constraints on the local driver mechanisms responsible for E s. It would also provide useful information for communication professionals using HF and lower VHF frequencies. [41] The mapping procedure we have demonstrated is indirect in that ionospheric modeling is required to distinguish normal E/F region signal propagation from E s propagation. The mapping observations are performed by a network of affordable software-controlled radio receivers, which constitute a Distributed Array of Small Instruments (DASI). The DASI concept was highlighted in the National Research Council s Decadal Research Strategy [National Research Council, 2002] as a necessary approach to solving many outstanding science questions in space physics. Extending the array of dedicated receivers and augmenting it with other propagation observations (e.g., PropNET) will lead to a cost-effective means of mapping E s creation, dynamics, and evolution. [42] The receiver network operates in real time, and the central site collects the observations and interprets results according to the modeled MUF. As described previously, the IRI model is currently used to estimate the MUF. However, with the availability of assimilation modeling, e.g., USU-GAIM [Schunk et al., 1997], higher-quality 10 of 11

11 estimates including current space weather effects obtained from ionosondes and GPS TEC might be incorporated by the central site. [43] The resulting E s maps would in turn be useful to assimilative models, providing information about observed MUFs and possibly allowing corrections to nighttime GPS TECs, where intense E s can have a small but measurable effect. Improved modeling, as well as the availability of E s maps, would be beneficial to both the communications and scientific communities. 7. Summary [44] The July 2008 SWARE campaign carried out in the western United States demonstrated a passive technique to map E s. Data analysis showed that E s was indeed responsible for the propagation of radio waves above the climatological MUF values used as thresholds. This achievement has the potential to lead to a larger-scale DASI that would produce scientifically valuable maps of E s over the CONUS region. The SWARE network would complement existing ionosonde observations: ionosondes provide detailed E s measurements at a given location, and a dense SWARE network could fill in the gaps between these widely spaced ionosonde values at a low cost and with no impact on the radio environment. [45] The E s observations made by the prototype SWARE network can provide regional warnings of the presence of E s. These warnings could be used by HF communicators as indicators that propagation prediction software may not provide reliable results and that unusual propagation may be observed. [46] As the observations are analyzed in more detail, it should be possible to study the evolution and morphology of E s through the solar cycle. It will be especially interesting to see how E s behavior changes as solar activity increases. Data from other sources, such as the USU meteor wind radar at BLO, may add valuable insight into the nature of the E s blooms. [47] Acknowledgments. This research was supported by contracts FA C-0043 and FA C-0016 from the Air Force Research Laboratory at Hanscom AFB to Space Environment Corporation. We also wish to acknowledge the contribution of many radio amateurs through the PropNET project ( References Arras, C., C. Jacobi, and J. Wickert (2009), Semidiurnal tidal signature in sporadic E occurrence rates derived from GPS radio occultation measurements at higher midlatitudes, Ann. Geophys., 27, , doi: /angeo Bilitza, D. (2001), International Reference Ionosphere 2000, Radio Sci., 36, , doi: /2000rs Christakis, N., C. Haldoupis, Q. Zhou, and C. Meek (2009), Seasonal variability and descent of mid-latitude sporadic E layers at Arecibo, Ann. Geophys., 27, , doi: /angeo Coleman, C. J. (1993), A General Purpose Ionospheric Ray Tracing Procedure, Tech. Rep. SRL-0131-TR, Def. Sci. and Technol. Organ., Salisbury, South Australia. David, M., and J. J. Sojka (2010), Single-day dayside density enhancements over Europe: A survey of a half-century of ionosonde data, J. Geophys. Res., 115, A12311, doi: /2010ja Eccles, J. V., R. D. Hunsucker, D. Rice, and J. J. Sojka (2005), Space weather effects on midlatitude HF propagation paths: Observations and a data-driven D-region model, Space Weather, 3, S01002, doi: /2004sw Eccles, J. V., D. D. Rice, J. J. Sojka, J. W. Raitt, and R. D. Hunsucker (2008), Frequency-agile distributed-sensor system (FADSS) deployment in the western United States: HF results, in 12th International Ionospheric Effects Symposium, edited by J. M. Goodman et al., pp , Natl. Tech. Inf. Serv., Alexandria, Va. Ford, S. (2008), PropNET, QST, 92(3), 95. Garcia-Fernandez, M., and T. Tsuda (2006), A global distribution of sporadic E events revealed by means of CHAMP-GPS occultations, Earth Planets Space, 58, Haldoupis, C., D. Pancheva, and N. J. Mitchell (2004), A study of tidal and planetary wave periodicities present in midlatitude sporadic E layers, J. Geophys. Res., 109, A02302, doi: /2003ja Hunsucker, R. D. (1991), Radio Techniques for Probing the Terrestrial Ionosphere, Phys. and Chem. in Space Ser., Vol. 22, Springer, New York. Mathews, J. D. (1998), Sporadic E: Current views and recent progress, J. Atmos. Sol. Terr. Phys., 60, , doi: /s (97) Miya, K. (1996), My memory of efforts in developing CCIR Recommendation 534 3: Method for calculating sporadic-e field strength, IEEE Antennas Propag. Mag., 38, 90 93, doi: / National Research Council (2002), The Sun to the Earth and Beyond: A Decadal Research Strategy in Solar and Space Physics, Natl. Acad. Press, Washington, D. C. Neubeck, K. E. (1996), Using the combined resources of amateur radio observations and ionosonde data in the study of temperate zone sporadic-e, J. Atmos. Terr. Phys., 58, , doi: / (95) Rice, D. D., J. V. Eccles, J. J. Sojka, J. W. Raitt, J. Brady, and R. D. Hunsucker (2009), A frequency agile distributed sensor system to address space weather effects upon ionospherically dependent systems, Radio Sci., 44, RS0A29, doi: /2008rs Schunk, R. W., and A. F. Nagy (2000), Ionospheres: Physics, Plasma Physics, and Chemistry, Atmos Space Sci. Ser., Cambridge Univ. Press, Cambridge, U. K. Schunk, R. W., J. J. Sojka, and J. V. Eccles (1997), Expanded Capabilities for the Ionospheric Forecast Model, Rep. AFRL-VS-HA-TR , Air Force Res. Lab., Hanscom AFB, Mass. Smith, E. K. (1962), The occurrence of sporadic E, Ionospheric Sporadic E, edited by E. K. Smith and S. Matsushita, pp. 3 12, Pergamon Press, New York. Smith, E. K. (1976), World Maps of Sporadic E (f o E s > 7 MHz) for Use in Prediction of VHF Oblique-Incidence Propagation, Off. Telecommun. Spec. Publ , U.S. Dep. of Commer., Boulder, Colo. Sojka, J. J., D. Rice, and J. V. Eccles (2005), Affordable software radio: A new tool for HF science and space weather operations, in 2005 Ionospheric Effects Symposium, edited by J. M. Goodman et al., pp , Natl. Tech. Inf. Serv., Alexandria, Va. Whitehead, J. D. (1989), Recent work on mid-latitude and equatorial sporadic-e, J. Atmos. Terr. Phys., 51, , doi: / (89) Wu, D. L., C. O. Ao, G. A. Hajj, M. de la Torre Juarez, and A. J. Mannucci (2005), Sporadic E morphology from GPS-CHAMP radio occultation, J. Geophys. Res., 110, A01306, doi: / 2004JA Yokoyama, T., T. Horinouchi, M. Yamamoto, and S. Fukao (2004), Modulation of the midlatitude ionospheric E region by atmospheric gravity waves through polarization electric field, J. Geophys. Res., 109, A12307, doi: /2004ja J. J. Brady, J. V. Eccles, J. W. Raitt, D. D. Rice, and J. J. Sojka, Space Environment Corporation, 221 N. Gateway Dr., Ste. A, Providence, UT , USA. (jan.sojka@usu.edu) R. D. Hunsucker, RP Consultants, 7927 Gearhart St., Klamath Falls, OR 97601, USA. 11 of 11