+ All Categories
Home > Documents > Oblique chirp sounding and modeling of ionospheric HF ...

Oblique chirp sounding and modeling of ionospheric HF ...

Date post: 02-Dec-2021
Category:
Upload: others
View: 4 times
Download: 0 times
Share this document with a friend
18
INTERNATIONAL JOURNAL OF GEOMAGNETISM AND AERONOMY VOL. 7, GI2002, doi:10.1029/2006GI000143, 2007 Oblique chirp sounding and modeling of ionospheric HF channel at paths of different length and orientation G. G. Vertogradov, 1 V. G. Vertogradov, 1 and V. P. Uryadov 2 Received 20 January 2006; accepted 12 February 2007; published 11 May 2007. [1] The results of systematic observations during 2005 of the main characteristics of HF propagation at a network of paths of oblique linear frequency modulation (LFM) sounding of the ionosphere are presented. The values of the maximum observed frequency (MOF) are calculated using the IRI 2001 model and the effective sunspot number W eff . It is shown that the difference between the predicted and measured monthly mean values does not exceed 12%. It is found that at all propagation paths, quasiperiodic variations of MOF are almost always observed. The amplitude of the MOF variations in the daytime can reach 2 MHz. A spectral analysis of the MOF variations at the midlatitude paths Cyprus– Rostov-on-Don and Inskip–Rostov-on-Don is performed. It is shown that the spectra of the MOF fluctuations have a well-pronounced linear structure. The power of spectral components of MOF fluctuations is concentrated in the 20–90 min range. Analyzing the dynamics of motion of z-type features on a trace of the high-angle ray, the periods of traveling ionospheric disturbances (TID) are determined and found to be 15–30 min. It is found that the TID with such periods correlate well with the sunset and sunrise for the middle point of the path. Modeling of oblique sounding (OS) ionograms at the presence of TID is performed taking into account the procedure of LFM signal processing. On the basis of the comparison of the experimental and calculated data, parameters of middle-scale wave disturbances responsible for the formation of z-type features in OS ionograms are determined. INDEX TERMS: 0689 Electromagnetics: Wave propagation; 2487 Ionosphere: Wave propagation; 2435 Ionosphere: Ionospheric disturbances; KEYWORDS: Ionospheric HF channel; Modeling of oblique sounding ionograms; Traveling ionospheric disturbances. Citation: Vertogradov, G. G., V. G. Vertogradov, and V. P. Uryadov (2007), Oblique chirp sounding and modeling of ionospheric HF channel at paths of different length and orientation, Int. J. Geomagn. Aeron., 7, GI2002, doi:10.1029/2006GI000143. 1. Introduction [2] The efficiency of communication systems in a signifi- cant degree depends on the coordination of the transmitted signals with the propagation channel. At the same time, the problems of coordination in the HF range can be resolved successfully only at understanding of the specifics of the im- pact of the propagation medium on the HF signals. Similar problems arise also at the development of models of the HF channel describing adequately the real features of HF prop- agation. [3] Finally, the adaptation of both the channel models and real HF communication systems to current ionospheric 1 Rostov State University, Rostov-on-Don, Russia 2 Radiophysical Research Institute, Nizhny Novgorod, Russia Copyright 2007 by the American Geophysical Union. 1524–4423/07/2006GI000143$18.00 conditions is possible only at the presence of radio channel sounding means with use of real-time data. Systems of both, pulse [Al’pert, 1972; Blagoveshchensky and Zherebtsov, 1987; Davies, 1969; Moller, 1974] and continuous [Filipp et al., 1991; Poole and Evans, 1985] vertical (VS), oblique (OS), and oblique backscatter (OBS) sounding are used for these purposes. The pulse methods, being simple and reliable, have, as it is widely known, one very important disadvan- tage: they require high power of emission in a pulse, the power seldom being below 1 kW. [4] On the basis of the sounding methods using continuous wide-band signals at much lower emitting power (usually 10 or 100 W), one can measure not only the traditional param- eters (the mode structure, relative delays, the ratio of ray amplitudes, the signal-to-noise ratio, etc.), but also quasi- instant transmission and pulse characteristics of ionospheric radio channel. The method of ionospheric channel diagnos- tics with the use of signals with the LFM, the so-called chirp sounding (R. B. Fenwick, Commun. News, 1974) is widely GI2002 1 of 18
Transcript
Page 1: Oblique chirp sounding and modeling of ionospheric HF ...

INTERNATIONAL JOURNAL OF GEOMAGNETISM AND AERONOMYVOL. 7, GI2002, doi:10.1029/2006GI000143, 2007

Oblique chirp sounding and modeling of ionospheric HFchannel at paths of different length and orientation

G. G. Vertogradov,1 V. G. Vertogradov,1 and V. P. Uryadov2

Received 20 January 2006; accepted 12 February 2007; published 11 May 2007.

[1] The results of systematic observations during 2005 of the main characteristics of HFpropagation at a network of paths of oblique linear frequency modulation (LFM) soundingof the ionosphere are presented. The values of the maximum observed frequency (MOF)are calculated using the IRI 2001 model and the effective sunspot number Weff . It is shownthat the difference between the predicted and measured monthly mean values does notexceed 12%. It is found that at all propagation paths, quasiperiodic variations of MOFare almost always observed. The amplitude of the MOF variations in the daytime canreach 2 MHz. A spectral analysis of the MOF variations at the midlatitude paths Cyprus–Rostov-on-Don and Inskip–Rostov-on-Don is performed. It is shown that the spectra ofthe MOF fluctuations have a well-pronounced linear structure. The power of spectralcomponents of MOF fluctuations is concentrated in the 20–90 min range. Analyzing thedynamics of motion of z-type features on a trace of the high-angle ray, the periods oftraveling ionospheric disturbances (TID) are determined and found to be 15–30 min. It isfound that the TID with such periods correlate well with the sunset and sunrise for themiddle point of the path. Modeling of oblique sounding (OS) ionograms at the presenceof TID is performed taking into account the procedure of LFM signal processing. On thebasis of the comparison of the experimental and calculated data, parameters of middle-scalewave disturbances responsible for the formation of z-type features in OS ionograms aredetermined. INDEX TERMS: 0689 Electromagnetics: Wave propagation; 2487 Ionosphere: Wave propagation;

2435 Ionosphere: Ionospheric disturbances; KEYWORDS: Ionospheric HF channel; Modeling of oblique sounding

ionograms; Traveling ionospheric disturbances.

Citation: Vertogradov, G. G., V. G. Vertogradov, and V. P. Uryadov (2007), Oblique chirp sounding and modeling of ionospheric

HF channel at paths of different length and orientation, Int. J. Geomagn. Aeron., 7, GI2002, doi:10.1029/2006GI000143.

1. Introduction

[2] The efficiency of communication systems in a signifi-cant degree depends on the coordination of the transmittedsignals with the propagation channel. At the same time, theproblems of coordination in the HF range can be resolvedsuccessfully only at understanding of the specifics of the im-pact of the propagation medium on the HF signals. Similarproblems arise also at the development of models of the HFchannel describing adequately the real features of HF prop-agation.

[3] Finally, the adaptation of both the channel modelsand real HF communication systems to current ionospheric

1Rostov State University, Rostov-on-Don, Russia2Radiophysical Research Institute, Nizhny Novgorod, Russia

Copyright 2007 by the American Geophysical Union.

1524–4423/07/2006GI000143$18.00

conditions is possible only at the presence of radio channelsounding means with use of real-time data. Systems of both,pulse [Al’pert, 1972; Blagoveshchensky and Zherebtsov, 1987;Davies, 1969; Moller, 1974] and continuous [Filipp et al.,1991; Poole and Evans, 1985] vertical (VS), oblique (OS),and oblique backscatter (OBS) sounding are used for thesepurposes. The pulse methods, being simple and reliable,have, as it is widely known, one very important disadvan-tage: they require high power of emission in a pulse, thepower seldom being below 1 kW.

[4] On the basis of the sounding methods using continuouswide-band signals at much lower emitting power (usually 10or 100 W), one can measure not only the traditional param-eters (the mode structure, relative delays, the ratio of rayamplitudes, the signal-to-noise ratio, etc.), but also quasi-instant transmission and pulse characteristics of ionosphericradio channel. The method of ionospheric channel diagnos-tics with the use of signals with the LFM, the so-called chirpsounding (R. B. Fenwick, Commun. News, 1974) is widely

GI2002 1 of 18

Page 2: Oblique chirp sounding and modeling of ionospheric HF ...

GI2002 vertogradov et al.: modeling of ionospheric hf channel GI2002

applied in practice of radio communication [Goodman, 1992].On the other hand, the use of small-power wide-band signalsat continuous sounding puts additional requirements both oncreation of the sounding equipment and on the choice anddevelopment of the means and methods of processing of theobtained information.

[5] It is also important to note that the developmentof wide-band adaptive communication and direction-findingsystems of the decameter range leads to an important roleof the problems of studies of dynamical spectral and sta-tistical characteristics of a nonstationary ionospheric radiochannel. The nonstationarity of ionospheric channel is in asignificant degree due to the influence of wave disturbancesof different origin generated at passage of terminator andduring geomagnetic disturbances.

[6] Medium-scale traveling ionospheric disturbances (TID)with dimensions of 100–500 km are typical ionospheric dis-turbances observed at middle latitudes [Danilov et al., 1987;Hocke and Schlegel, 1996]. Different methods and techniquesare used for studying TID, including ionosondes of verti-cal [Bowman, 1990; MacDougall et al., 1997] and oblique[Cherkashin et al., 2003] sounding, Doppler measurements[Burmaka et al., 2003; Waldock and Jones, 1987], incoher-ent scatter radars [Burmaka et al., 2004; Huang et al., 2003;Hearn and Yeh, 1978], MU radars [Fukao et al, 1991], andtransionospheric sounding with the help of signals of theGPS navigation satellites [Ho et al., 1998]. Beginning fromthe second half of the 1990s, optical observations of thenightglow of the ionosphere are widely used [Shiokawa etal., 2003; Taylor et al., 1998].

[7] In spite of considerable efforts and achievements in thefield of wave disturbance studies, many important questionsstill are open. The mechanism of transformation of varioustypes of energy in the atmosphere is still obscure; the agentstimulating the trigger mechanism of the accumulated en-ergy release is not still found. The patchiness of the dataand their discrepancy lead in some cases to erroneous con-clusions.

[8] Because of the complexity and variety of relations inthe solar wind-magnetosphere-ionosphere-atmosphere-Earthsystem and also the presence of various physical mechanismsresponsible for generation of wave disturbances and theirnonstationarity, carrying out of coordinated systematic stud-ies of wave disturbances seems very actual. The entire setof the observation methods and equipments available shouldbe attracted to such studies. In our mind, the network ofthe stations of oblique LFM sounding operating on the reg-ular basis in automatic regime is one of effective methods oforganizing systematic observations for manifestation of wavedisturbances.

[9] In this paper, the results of regular observations of theconditions of ionospheric propagation of HF radio waves atradio paths of different length and orientation obtained withthe help of the network of chirp sounders are presented. Themeasurements were conducted in 2005. Vast many-monthcontinuous set of data on the frequency-time and frequency-amplitude display ionograms was obtained. The registra-tion and processing of the data were performed in an auto-matic regime. It is shown that the method of oblique LFMsounding with high resolution in the group delay time and

frequency makes it possible to detect reliably wave distur-bances with a period from 15 min to a few hours and withan amplitude of variations of the MOF from 0.1 MHz andhigher.

[10] Spectral analysis of variations in MOF and the effectof appearance of z-type features at the high-angle ray ofionograms of the oblique sounding are used for revealing ofwave disturbances. The high-angle ray slides along the layermaximum and is considered as an unstable trajectory. Thehigh-angle ray is a sensitive detector of various disturbancesand due to this property is quite convenient for detecting ofionospheric wave disturbances [Erukhimov et al., 1997].

[11] The efficiency of using of the high-angle ray for detect-ing disturbances in a significant degree is determined by theavailability of modern equipment for wide-band sounding ofthe ionosphere with high resolution in frequency and groupdelay time, as well as by proper choice of the path geome-try. From the point of view of the sounding by modernizedautomats the LFM ionosonde modernized by the authorssatisfies completely the above indicated requirements. Suchfactors as the length and the orientation of the path influ-ence the detection of the wave disturbances by the high-angleray. At very long paths (multihop), the effect of disturbanceimpact on the characteristics of the high-angle ray is of anintegral character, this fact making difficult identification ofthe disturbance source. Moreover, at such paths the effect ofthe impact of ionospheric disturbances on signal character-istics may be masked by scattering of radio waves from theEarth. Scattering of radio waves by intense irregularities (ifthe path passes in the high-latitude ionosphere) can lead tothe same effect. On the other hand, at short paths with thelength ≤ 500–700 km the frequency range of the propagat-ing HF waves shifts toward lower frequencies, where the highlevel of stationary noise exists. This circumstance causes anincrease of the threshold of the TIDs detection. Startingfrom these requirements, the most suitable for revealing ef-fects of the impact of wave disturbances on the character-istics of the high-angle ray are midlatitude one-hop paths∼1500–3000 km long.

[12] In this paper a modeling of the frequency-time andfrequency-amplitude display ionograms is performed forboth, conditions of quiet ionosphere and at the presenceof wave disturbances. Comparing the experimental andcalculated ionograms of oblique sounding, we determinedthe parameters of wave disturbances causing such featureson ionograms as z-type disturbances of the high-angle ray.

2. Equipment and the Method of DataProcessing

[13] The measurements were conducted at the radio pathsof LFM sounding of different length and orientation: Cyprus(35◦N, 34◦E)–Rostov-on-Don (47.2◦N, 39.7◦E) (the pathlength is 1440 km, the geographical azimuth from Rostov-on-Don is 203.2◦), Inskip (England, 53.8◦N, 2.8◦W)–Rostov-on-Don (3043 km, 300◦), Norilsk (69.4◦N, 88.2◦E)–Rostov-on-Don (3587 km, 29.8◦), Irkutsk (52.3◦N, 104.2◦E)–Rostov-on-

2 of 18

Page 3: Oblique chirp sounding and modeling of ionospheric HF ...

GI2002 vertogradov et al.: modeling of ionospheric hf channel GI2002

Figure 1. Geometry of the paths during experiments.

Don (4512 km, 57.8◦), Magadan (59.6◦N, 150.8◦E)–Rostov-on-Don (6615 km, 33.4◦). At two former paths the mea-surements were carried out during the entire 2005 aroundthe clock with one ionogram recorded every 5 min. At thelatter three paths the measurements were carried out duringSeptember 2005 around the clock with the interval of oneionogram recording of 15 min at each path. The receptionpoint was located at Rostov-on-Don. Geometry of paths isshown in Figure 1.

[14] Two-channel chirp sounder created on the basis of the“Katran” R-399 A receiver was used in the measurements.The rate of the frequency sweeping was 100 kHz s−1. Therange of emitting frequencies was 8–30, 4.2–30, and 4–30MHz for Cyprus, Inskip, and the rest of the transmitters,respectively. Two 9-m collapsible-whip antennas were usedfor the reception. The time synchronization of the start ofreception of LFM signals was provided with the help of GPSwith the error less than 10 µs. The residual signal was dig-itized at intermediate frequency (IF) (f = 215 kHz) usingthe 14-digit analog-digital transformer (ADT) with the fre-quency discreteness of 50000 Hz, the latter value exceedingconsiderably the used transmission band of the receiver atIF (3000 Hz).

[15] The signal transformation and its processing in-cluded the following stages. The received residual signalwas put through the procedures of a transfer to the zerofrequency with obtaining quadrature components (complexlow-frequency envelope), low-frequency filtration by a digitalfilter with the transmission band of 500 Hz, and decimationwith the reduction of the discreteness frequency down to

3000 Hz. As a result, the procedure described increased thedynamical range not less than by 10 dB. The proceduresof digitalization, filtration, signal quadrature revealing, anddecimation were organized in such a way that the entirepreliminary processing was performed in real time in anautomatic way and made it possible to obtain continuousrecordings of unlimited duration in time. Further spectralprocessing of the differential signal was performed based onmultitaper method [Thomson, 1982] with the aim to extractcontinuous and discrete components of spectral density ofpower. The determination of ray parameters (the amountof rays, their amplitudes and time delays) is based on theestimation of the discrete components in the spectrum of theresidual signal. Such estimate is checked up based on thethreshold statistical criterion (statistics of F distribution)[Thomson, 1982]. Noise spectral density and the ratio ofthe signal power to the noise power in the reception bandconsistent with the signal are determined by the histogrammethod assuming that the frequency band of the spectralanalysis is much larger than the frequency band of the re-ceiving signal. For such purpose the histogram of the powerspectral density is obtained in the frequency band of thereceiver. The maximal level of histogram corresponds tothe probable value of the noise spectral density during theperiod of the data obtaining. After that the ratio of the sig-nal power to the noise power is determined quite evidently.As a result of the processing in real time, the followingparameters were determined: the level of the noise spectraldensity in the reception band, the number of detected prop-agation rays, amplitudes of all rays, the signal-to-noise ratio

3 of 18

Page 4: Oblique chirp sounding and modeling of ionospheric HF ...

GI2002 vertogradov et al.: modeling of ionospheric hf channel GI2002

for each ray, and absolute delays of each of detected rays[Vertogradov, 2005].

[16] Further processing had the aim to obtain the freq-uency-time and frequency-amplitude display ionograms.Frequency-delay and frequency-amplitude display of thesingle propagating rays and modes are formed by compar-ison of the obtained parameters in the multivariate spacefrequency-time-amplitude. Each point that at the beginningwas identified with some ray is characterized by three coor-dinates: fj – frequency, τjl – time delay, ajl – amplitude ofthe lth ray on the jth frequency. The amount of rays, nj ,and their coordinates in the three-dimensional (3-D) spaceare determined according to the algorithm described abovein the process of spectral processing based on multiple-tapermethod spectral analysis. In the 3-D space frequency-time-amplitude we introduce the distance between pointsaccording to the rule

ρj,l;i,m = ρ(fj , τjl, ajl; fi, τim, ail) =

|fj − fi|εf

+|τjl − τil|

ετ+|ajl − ail|

εa.

[17] Two points that correspond to the neighbor frequen-cies fj−1 and fj with the indices l and m are assumedto belong to the same frequency branch if the relationρj,l;j−1,m < 3 is realized. The parameters εf , ετ , and εa areselected empirically and assumed to be 500 kHz, 0.1 ms, and3 dB correspondingly. If for some point the correspondencewith the previous points is not established, then we assumethat this point relates to a new branch. As a consequence atthe end of the process of sounding many frequency branchesare formed. After that the procedure of the secondary pro-cessing takes place. At this stage the regions restricted bythe lowest (LOF) and the maximal (MOF) observed frequen-cies are found for each branch. The total amount of theseregions for all frequency branches provides the opportunityto determine the intervals: the frequency intervals whereone ray, two rays, three rays, etc., exist. From the MOF forall rays the MOF of the path is determined as the maximalfrequency from all observed frequencies. From the lowest fre-quencies the LOF of the path is determined as the minimalfrequency from all observed frequencies.

[18] The band of the analysis of the residual signal was65 kHz, the latter value providing the separation of modesand propagation rays with a resolution better than 15 µs.The spectral processing of the residual signal was performedwith the time window equal to 650 ms. At each step it wasshifted at 1/16 part. So the points in the space frequency-time-amplitude are separated by ∼4 kHz. As a consequence,the accuracy with which the frequency boundaries of multi-ray region and the tool accuracy of the LOF and the MOFestimate comprises ∼4 kHz. As the signal is the diffuseone, the real accuracy of the LOF and the MOF estimatecomprises ∼10–20 kHz. Examples of frequency-time andfrequency-amplitude display for the Cyprus–Rostov-on-Donpath (where the effects of wave disturbances are manifestedin the most visual way) are shown in Figures 2 and 3. Fig-ure 2 shows the dynamics of transportation of a wave dis-

turbance (z-type features) at the track of the high-angle rayfor 1F mode. Figure 3a shows an example of registration ofseveral high-angle rays accompanied by the F spread phe-nomenon. Figure 3b illustrates the effect of quasi-regularamplitude-frequency modulation of the high-angle ray. Theeffect is well seen at frequency time display for the trackof extraordinary mode of high-angle ray (Figure 3b, top),and also at frequency-amplitude display (Figure 3b, bottom)for the ordinary and extraordinary modes of high-angle ray(Pedersen mode), marked as 1Fpo and 1Fpx, respectively.Figures 3c and 3d show examples of registration of iono-spheric disturbances simultaneously at modes 1F and 2F.In the procedure of processing, the amplitude display werereduced to the maximum digitalization of the amplitude-digital transformer (ADT), in other words, 0 dB correspondsto digitalization of the sinusoidal signal with the amplitudeequal to the maximum digits of ADT. It is worth noting alsothat the amplitude of separated rays was determined by in-tegrating of the spectral density of the residual signal overthe vicinity of the maximums in the spectrum limited by thenearest maximums from the right-hand and left-hand sides.This operation made it possible to reach much higher accu-racy of measuring the amplitudes of partial rays than justuse of the amplitude value in the maximum of the spectraldensity.

[19] The further processing of frequency-time and freq-uency-amplitude display ionograms at the studied paths wasaimed at (1) obtaining of the diurnal variations of the max-imum observed frequency of particular propagation modes,MOF was determined for extraordinary modes; (2) obtainingof averaged monthly mean variations in MOF of particularpropagation modes; (3) comparison of the averaged monthlymean values of MOF with the forecasted values based onthe International Reference Ionosphere IRI 2001; (4) spec-tral analysis of the time variations in MOF; (5) analysis ofthe frequency-time display and its variations caused by TID;and (6) analysis of the frequency-time display for particularpropagation modes and comparison of the experimental de-pendencies with the results of modeling.

3. Observational Results

[20] Figure 4 shows the results of the processing of obliquesounding ionograms obtained in 2005 at paths of variouslength and orientation. Each picture corresponds to monthlycontinuous observations and contains data for not less than6000 sounding sessions. Using the algorithm developed byVertogradov [2005], the values of MOF for each session weredetermined automatically in real time. Thick line in theplots presents the result of the monthly statistical averag-ing. Together with estimation of the mathematical expecta-tion, the estimate of the standard mean deviation from themathematical expectation was found. Thin dashed lines inthe plots limit the band in the vicinity of the mathematicalexpectation, the width of the band being equal to the stan-dard deviation doubled. The forecasted value of the max-imum usable frequency (MUF) for the corresponding pathwas calculated for the middle of each month. The MUF

4 of 18

Page 5: Oblique chirp sounding and modeling of ionospheric HF ...

GI2002 vertogradov et al.: modeling of ionospheric hf channel GI2002

Figure 2. Dynamics of variations in frequency-time display ionograms (the top part of each pairsof the plots) and frequency-amplitude display ionograms (the bottom part of each pair of the plots)in the presence of TID at the sounding path. Amplitude display corresponding to the ordinary andextraordinary modes and the Pedersen ray (high-angle ray) are marked by o, x, and p indices.

5 of 18

Page 6: Oblique chirp sounding and modeling of ionospheric HF ...

GI2002 vertogradov et al.: modeling of ionospheric hf channel GI2002

Figure 3. Examples of frequency-time display ionograms (the top part of each pairs of the plots) andfrequency-amplitude display ionograms (the bottom part of each pair of the plots) in the presence of TIDat the sounding path.

6 of 18

Page 7: Oblique chirp sounding and modeling of ionospheric HF ...

GI2002 vertogradov et al.: modeling of ionospheric hf channel GI2002

was calculated by the x mode. Simultaneously the modelIRI-2001 was corrected by the index of solar activity. Theeffective values of Weff were used as the index of solar activ-ity [Secan and Wilkinson, 1997]. The values of Weff for eachmonth are shown in Figure 4. The forecasted values of MUFobtained in the above described way are shown in Figure 4by crosses.

[21] On the basis of the processing of continuous diurnaldependencies of MUF obtained in 2005, one can draw thefollowing conclusions.

[22] The MOF values for some propagation modes undergoshort periodic variations with quasi-periods of ∼20 min to1.5 hours. In quiet ionospheric conditions in the afternoonhours of the day, the fluctuation amplitude can reach 2 MHz.In separate days in dawn-dusk hours at passing of the termi-nator, the fluctuations at the Cyprus–Rostov-on-Don pathcan grow up to 5–8 MHz.

[23] At monthly averaging the standard deviations (SD)of MOF from the mathematical expectation at paths of alllengths are higher in the daytime and increase with a pathlength increase. The typical SDs of MOF in the daytime forvarious paths have the following values: 1.2 MHz (Cyprus–Rostov-on-Don), 1.7 MHz (Inskip–Rostov-on-Don), 2.0 MHz(Norilsk–Rostov-on-Don), 2.2 MHz (Irkutsk– Rostov-on-Don), and 2.5 MHz (Magadan–Rostov-on-Don). At nightSD of MOF decreases (very weak dependence on thepath length is still left) down to the following values: 1.0MHz (Cyprus–Rostov-on-Don), 1.5 MHz (Inskip–Rostov-on-Don), 1.5 MHz (Norilsk–Rostov-on-Don), 1.5 MHz(Irkutsk–Rostov-on-Don), and 1.6‘MHz (Magadan–Rostov-on-Don). No seasonal dependence of SD of MOF was foundon the basis of the measurements at the Cyprus–Rostov-on-Don and Inskip–Rostov-on-Don paths. Thus the meandiurnal variations of MOF about 1.5–2.5 MHz in the vicinityof the maximum values of MOF (in the daytime) and about1.0–1.6 MHz in the vicinity of the minimum values of MOF(at night) should be taken as typical values for midlatitudepaths of various length in quiet geophysical conditions.

[24] As it has been already mentioned above, MOF of the1F mode undergo quasiperiodical variations observed almostpermanently at various paths. At long multihop paths thiseffect is partly masked due to high diffusity of the signalcaused by the reflection from the Earth and the scatteringin the ionosphere. The most distinct variations in MOFwere observed at the Cyprus–Rostov-on-Don path 1440 kmlong. We performed an analysis of these variations aimedat detecting the spectral composition of MOF fluctuations.Two approaches were applied to get rid of the diurnal trendin the time series of MOF. In the first approach, the trendwas found by a running averaging at the chosen width ofthe time window ∆T . Then the trend was withdrawn fromthe initial time series of MOF. Such procedure is equivalentto high-frequency filtration. It is not optimal due to highsidelobes of the digital filter. As a consequence the timeseries of MOF fluctuations save traces with quasi-periodsexceeding considerably the width of the time window. Inthe second approach, the trend was withdrawn on the basisof a digital high-frequency (HF) filtration of the initial seriesof MOF. The frequency of the filter cutoff was chosen equalto F = 1/∆T . The tangent Butterworth filter of the tenth-

order was used as the digital filter. This procedure providedeffective suppression of the components of the diurnal vari-ations in MOF with frequencies below the cutoff frequencyF of the digital filter.

[25] The spectral analysis was performed using the fastFourier transform (FFT) algorithm. The examples of thespectral density for the Cyprus–Rostov-on-Don path for aseries of consequent days of various months are presented inthe form of sonogram in Figure 5. Analyzing the graphs ofthe spectral density of the power of MOF fluctuations, onecan conclude the following:

[26] 1. For different seasons the spectra of the MOF fluc-tuations have a well-pronounced linear structure, this factmanifesting that the MOF fluctuations are due to propaga-tion in the ionospheric plasma of a train of quasi-harmonicwaves.

[27] 2. The power of spectral components of MOF fluctu-ations is concentrated in the 20–90 min range.

[28] 3. The spectral composition of the MOF fluctuationsvaries from one day to another, though some quasi-harmoniccomponents can exists in MOF fluctuation spectra during afew days.

[29] In order to reveal the main quasi-periods of MOFfluctuations, the results of the spectral processing were sub-jected to a statistical analysis. To do that, in each spectrum(for a day) 5 the most intense harmonics were chosen andused to determine the corresponding quasi-periods. Then,using the monthly data obtained, the histograms of distri-bution of quasi-periods at the Cyprus–Rostov-on-Don andInskip–Rostov-on-Don paths were drawn. The histogramsare shown in Figure 6.

[30] The results obtained make it possible to draw thefollowing conclusions.

[31] 1. The most probable quasi-periods for both pathsare concentrated within the 20–90 min interval.

[32] 2. The positions of maximums and their numberchange from one month to another.

[33] 3. The most high-frequency component in the fluctu-ation spectrum has a quasi-period in the vicinity of 15 minand is present at both paths in all seasons, though the am-plitude of this quasi-harmonics varies from one month toanother.

[34] Fluctuations of MOF are accompanied by appearancethe z-type features in frequency-time display ionograms (seeFigures 2 and 3) what appear in the vicinity of the low-est observed frequency (LOF) for the high-angle rays anddrift with time to the region of lower delays (into the vicin-ity of MOF). These features were observed in various timeof the day at all paths where the observations were con-ducted. However, most often and with maximum ampli-tudes, the z-type features were registered at the one-hoppaths Cyprus–Rostov-on-Don and Inskip–Rostov-on-Don atdawn-dusk hours of the day at the passage of the terminator.Typical amplitudes of z-type features at frequency-time dis-play achieve values ∼0.1–0.5 MHz. On the basis of modeling,we will show below that the z-type features are due to themotion of traveling ionospheric disturbances (TID). At loweramplitudes of TID, breaks and steps of various amplitudesare observed in frequency-time display. We analyzed the ap-pearance of TID detected by means of the high-angle ray

7 of 18

Page 8: Oblique chirp sounding and modeling of ionospheric HF ...

GI2002 vertogradov et al.: modeling of ionospheric hf channel GI2002

Figure 4. Diurnal behavior of MOF at the oblique sounding paths: thick line, crosses, and dashed linecorrespond to the experiment, the calculations using the IRI 2001 model with correction of the sunspotnumber Weff , and standard deviations of MOF, respectively.

8 of 18

Page 9: Oblique chirp sounding and modeling of ionospheric HF ...

GI2002 vertogradov et al.: modeling of ionospheric hf channel GI2002

Figure 4. (Continued).

9 of 18

Page 10: Oblique chirp sounding and modeling of ionospheric HF ...

GI2002 vertogradov et al.: modeling of ionospheric hf channel GI2002

Figure 5. Sonograms of the spectral density of the intensity of MOF variations at the Cyprus–Rostov-on-Don path for a series of consequent days of various months: (a) January 2005, (b) March 2005, (c)May 2005, and (d) October 2005.

10 of 18

Page 11: Oblique chirp sounding and modeling of ionospheric HF ...

GI2002 vertogradov et al.: modeling of ionospheric hf channel GI2002

Figure 6. Histograms of distribution of the periods of wave disturbances at the paths (left) Cyprus–Rostov-on-Don and (right) Inskip–Rostov-on-Don for various months of 2005 under the time smoothingwindow of 5000 s.

11 of 18

Page 12: Oblique chirp sounding and modeling of ionospheric HF ...

GI2002 vertogradov et al.: modeling of ionospheric hf channel GI2002

Figure 7. Probability of appearance of wave disturbanceswith periods of 15–30 min at the Cyprus–Rostov-on-Donpaths for (a) February, (b) March, and (c) April 2005. Ver-tical dashed lines show the moments of sunrise and sunsetfor the middle point of the path for the 15th day of eachmonth.

and found their period for the Cyprus–Rostov-on-Don pathwhere this effect was manifested most visually. The period ofTID responsible for formation of z-type feature was detectedfrom the frequency-time display ionogram analyzing the timeinterval between appearance at low frequencies of the high-angle ray of z-type feature, their drift along the track fromlow to high frequencies, and appearance of the wave trainat low frequencies immediately after disappearance of theprevious train of wave disturbances at high frequencies inthe vicinity of MOF. If on the track of the high-angle ray

one more z-type perturbation appears, this causes the un-certainty in the TID period determination. We exclude suchcases from our consideration. The detected in such a waytypical values of the TID periods were ∼15–30 min.

[35] Figure 7 shows a histogram of the distribution of prob-ability of TID appearance with time of the day for variousmonths of 2005 according to the data of the oblique LFMsounding of the ionosphere at the Cyprus–Rostov-on-Donpath. Vertical dashed lines show the sunrise and sunsetfor the middle point of the path for the 15th day of eachmonth. Different time interval of signal registration in dif-ferent months is due to the fact that the LFM transmitterat Cyprus starts from the frequency of 8 MHz which exceedsMOF at the Cyprus–Rostov-on-Don path at near-midnighthours for February (March). One can see in Figure 7 thatthe appearance of TID with periods 15–30 min correlateswell with the dawn and dusk for the middle point of thepath. At dawn-dusk hours the probability of TID appear-ance reaches 80–100%. These data confirm the importantrole of terminator as a source of TID generation.

[36] The traces of one-hop propagation modes (the Cyprus–Rostov-on-Don and Inskip–Rostov-on-Don paths), as a rule,have no indications to diffuse reflections and look like thinlines, their thickness being comparable with the Rayleighlimit of the spectral analysis. The traces of multiple modeshave some indications to diffuse reflections. The latter factis, evidently, due to the scattering at the reflection from un-even surface of the Earth. This statement is confirmed bythe observations at the Cyprus–Rostov-on-Don path. Theregion of the arrival of the first hope for the two-hop modefalls there at the sea-land boundary (the Black Sea coast ofTurkey). As a rule, the 2F mode at this path demonstrateswell-pronounced traces of diffusivity (see Figures 2 and 3b),this fact being, evidently, due to scattered reflection fromthe mountain Earth surface. At the same time, sometimes aclear trace of the 2F mode appears for a short period (see, forexample, Figures 3c and 3d). This trace looks like the one forthe 1F mode. The latter fact is, apparently, a result of thereflection from the sea surface. This event was observed atneither of the paths Inskip–Rostov-on-Don, Irkutsk–Rostov-on-Don, Norilsk–Rostov-on-Don, and Magadan–Rostov-on-Don. The multiple modes at these paths always have signsof diffuse reflections, the diffusivity degree increasing withan increase of the mode order.

[37] The frequency-amplitude display of some propaga-tion modes show deep fluctuations (up to 20–30 dB, see Fig-ures 2 and 3). The quasi-period and fluctuations depth varywith the frequency. Generation of the fluctuations can beexplained by interference of nonseparated rays within onepropagation mode. In this case the depth of the fluctuationsand quasi-period are determined by the relation between theray amplitudes and the difference in the group delays, respec-tively. It is known that, with the accuracy to linear terms,the signal phase may be presented in the form

φ(t, ω) = φ0 +∂φ

∂t∆t +

∂φ

∂ω∆ω = φ0 − 2πδf∆t + τ∆ω, (1)

where φ0 is the initial phase, δf is the Doppler shift of thefrequency, and τ is the group delay. One can see that the

12 of 18

Page 13: Oblique chirp sounding and modeling of ionospheric HF ...

GI2002 vertogradov et al.: modeling of ionospheric hf channel GI2002

quasi-period of amplitude fluctuations is determined by thedifference in the group delays of the nonseparated partialrays. Evidently, the zero beats will be observed, that is thefluctuation period will increase strongly, where the groupdelays of the rays coincide. This is observed in the region ofthe crossing of two magneto-ionic components at frequency-time display (see Figures 2 and 3).

[38] In the region where the magneto-ionic componentsare separated and only the trace of the extraordinary waveis left, the interference fluctuations at frequency-amplitudedisplay almost disappear (one-ray propagation).

[39] In the vicinity of the dead zone at frequency-amplitudedisplay of partial rays an effect of focusing is observed. Thevalue of the focusing can reach 10 dB. Similar effect takesplace in the vicinity of the “noses” of z-type features atfrequency-time display.

[40] This regular behavior of frequency-amplitude displayof particular propagation modes and rays can serve as a proofof the dominating propagation mechanism in the form ofdiscrete rays (mirror components). A casual input into theinitial phase in (1) due to small-scale structure of the iono-sphere appears to be small and insignificant. In the oppositecase, the random fluctuations at the amplitude display wouldhave been observed. Just so looks the amplitude display ofmultiple propagation modes.

4. Modeling and Discussion

[41] For interpretation of the obtained experimental re-sults and confirmation of the assumptions presented, we per-formed imitation modeling on the basis of the ray tracingwith the use of the imitation model of a wide-band iono-spheric radio channel [Barabashov and Vertogradov, 1996;Vertogradov, 2003, 2004]. The modeling is based on the so-lution of the expanded system of characteristic equations inthe ionosphere taken from the International Reference Iono-sphere IRI 2001 [Bilitza, 2001, 2002]. The ionospheric col-lision losses were calculated on the basis of the Appletonapproximation with the effective collision frequency foundon the basis of the values of the ion and electron concen-trations given by IRI 2001 with attraction of the neutralatmosphere model MSIS 90 with the utilization of knownratio [Gershman et al., 1984].

[42] The calculations of MUF of particular propagationmodes have shown that the IRI 2001 model, as a rule, givesoverestimated results, the discrepancy grows with the in-crease of the length path. Thus the IRI-2001 model needsan adaptation on the basis of the current geophysical infor-mation. Note that the IRI-2001 model assumes correctionwith respect to the sunspot numbers and the index of solaractivity. The correction of the solar activity index influencesonly on the F region of the ionosphere and does not affect thelower ionosphere which determines the main losses caused bycollisions of HF radio waves. As s consequence to providethe adaptation of the model to the experimental MOF valuesit is required only to correct the index of solar activity. Asgeneral geophysical information for correction of solar activ-ity index we took the data on the effective sunspot number

Weff [Secan and Wilkinson, 1997]. The results of the mod-eling showed that in this case one can obtain the predictedMUF values considerably closer to the measured values MOFat paths of all lengths and orientations for all seasons andsolar activity levels considered (Figure 4, crosses). The lat-ter result makes it possible to recommend using of predictedor current value of Weff as an index of solar activity in long-term or short-term forecasting of propagation characteristicson the basis of the IRI-2001 model. In that case the differ-ence between the forecasted and monthly mean measuredvalues does not exceed 12%. It is worth noting also thatafter the operation of correction of the IRI-2001 model bysolar activity index, the forecasted values, as a rule, fall intothe confidence interval with the width equal to the doubleroot-mean-square value of daily changes of MOF.

[43] For interpretation of the observed features of the high-angle ray, ionograms were calculated taking into accountTID [Vertogradov, 2003, 2004]. TID were modeled by mod-ulation of the average electron density N0(φ, θ, r, t) by a har-monic wave [MacDougall et al., 2001; Stocker et al., 2000].The instant electron density in the disturbed ionosphere atthe point with spherical coordinates φ, θ and r (r is the dis-tance from the center of the Earth, θ is counted from theaxis passing the North Pole, and φ is the latitude) at themoment of time t is written in the form

N(φ, θ, r, t) = N0(φ, θ, r, t)(1 + δN cos(kγ∆r+

kθr0∆θ + kφr0 sin θ∆φ− 2π

Tt + Φ0)), (2)

where r0 is the Earth radius, δN is the relative amplitudeof the disturbance, T is the TID period, Φ0 is the initialphase, {kφ = (2π/Λ) cos β sin α, kθ = (2π/Λ) cos β cos α,kγ = |2π/Λ| sin β} is the wave vector of TID with the wave-length Λ, α, and β provide the direction of the phase velocityof the wave disturbance, and ∆φ and ∆θ are the changes ofthe spherical coordinates relative to the emitting point.

[44] The modeling of oblique LFM sounding of the iono-sphere was performed using (2) in the situation in a maxi-mum degree close both to the solar and geophysical condi-tions of HF propagation at real path and to the methods ofprocessing of the residual signal. To do that, the structure-physical model of the HF radio channel was completed witha computer imitator of the wideband radio channel [Ver-togradov, 2003, 2004]. The modeling was performed in threestages.

[45] At the first stage, the radio channel is modeled for thegiven solar and geophysical conditions and the most impor-tant parameters of the complex transmission function of theHF channel in the given frequency band and in the giventime interval of the modeling are calculated. In our casethe important parameters of the transmission function werecalculated within the 4–32 MHz band, the step of the param-eters calculation being 0.05 MHz. The time interval of themodeling was chosen equal to 1800 s, the principal param-eters being calculated with a step of 10 s within the entireband of modeling.

[46] At the second stage on the basis of the principal pa-rameters of the HF channel, the complex transmission func-tion of the HF channel in the 4–32 MHz band continuous

13 of 18

Page 14: Oblique chirp sounding and modeling of ionospheric HF ...

GI2002 vertogradov et al.: modeling of ionospheric hf channel GI2002

in frequency and time is restored in the time interval 1800s. A wideband digital signal supplied to the input of thecomputer imitator of the HF channel is brought through thechannel with the found complex transmission function. Inour case the LFM signal was supplied to the imitator input,the rate frequency sweeping of the signal being chosen equalto 100 kHz s−1. At the output of the computer imitator,a low-frequency complex residual signal similar in its struc-ture to the real signal at the output of the LFM receiver wasobtained. The signal from the imitator output was saved inthe file. The file structure corresponded to the program ofprocessing of real residual signal saved in the process of realLFM sounding.

[47] At the third stage a processing of the recorded resid-ual signal by the program of processing of the signal at theoutput of a real LFM receiver was performed. The modeledresidual signal was processed by the program of spectral pro-cessing of residual signal for obtaining of frequency-time andfrequency-amplitude display ionograms.

[48] As a result, the procedure described makes it possibleto make the modeling stage in the maximum degree similarto the conditions of the reception and processing of a LFMsignal at real paths of the oblique sounding. In particular,the proposed algorithm of imitation modeling makes it pos-sible to take into account peculiarities related to the factthat frequency-time and frequency-amplitude display iono-grams at a real OS path are obtained not instantly but forthe sounding interval of about 300 s. The channel prop-erties can be changed during such time interval. Thus themain aim of the complete scheme of the imitation modelingcorresponding to the conditions of processing of a complex(linearly frequency-modulated) signal is to take into accountdynamical processes in the ionosphere during the soundingsession and to reveal effects related to the processing of theLFM signal at the presence of wave disturbances.

[49] The results of the imitation LFM sounding accordingto the complete scheme under various parameters of TID arepresented in Figures 8 and 9. On the basis of the imitationmodeling, one can draw the following conclusions:

[50] 1. Ionograms obtained on the basis of the model resid-ual signal agree qualitatively to the ionograms obtained inthe process of real oblique sounding of the ionosphere (seeFigures 2 and 3 and Figures 8 and 9). In particular, at thepresence of TID, the effect of amplitude-frequency modula-tion of the high-angle ray is observed in the model ionogramscalculated with the complete scheme (see Figure 9d) in thesame way as in the experiment (see Figure 3b (top), x modefor high-angle ray). In the same time, simple modeling ofthe ionospheric HF propagation does not reveal this effect.This fact should be taken into account at interpretation ofexperimental data.

[51] 2. A good agreement of the experimental and calcu-lated ionograms from the point of view of the TID influenceon both, the 1F and 2F modes is obtained (see Figures 3c,3d, 8a, 8d, 9a, 9d, and 9f).

[52] 3. The z-type features at frequency-time displayare actually caused by the TID motion in the ionosphericplasma. Taking into account of the finite time of the sound-ing does not change this conclusion.

[53] 4. The modeling showed that the conditions of forma-

tion of z-type features at frequency-time display are toughenough and depend on the relative amplitude of TID, thewavelength of the disturbance, and the direction of its prop-agation relative to the radio-path orientation. Such tracesin frequency-time display appear only under particular rela-tions between the amplitude and wavelength of TID undernot very small TID propagation angles to the horizon. Forexample, for the cases presented in Figure 8, (a–c) the TIDparameters were the following: the relative electron concen-tration δN = 20%, the wavelength Λ = 150 km, the pe-riod T = 15 min, the angle from the horizon β = −30◦.Nevertheless, during the whole 15 min period of modeling(3 ionograms) for 1F mode no z-type disturbances with ap-preciable amplitude were observed. At the same time theincrease of the angle β up to −60◦ caused the appearance ofthe z-type features at frequency-time display (see Figure 8,(d–f)). As a result, attraction of model calculations makesit possible to estimate TID parameters on the basis of theresults of oblique LFM sounding of the ionosphere. It fol-lows from the calculated ionograms that the z-type featuresare formed only under relative amplitudes δN of medium-scale disturbances exceeding 15% and under the TID prop-agation angles to the horizon plane from −60◦ to −30◦.In other words, for appearance of characteristic z-type fea-tures in oblique sounding ionograms, the vertical scale ofthe wave disturbance should not cover the entire F regionof the ionosphere. In the opposite case, the propagatingwave disturbance would lead only to quasiperiodic time vari-ations of MOF of the path, but not to local formations atthe frequency-time ionogram.

[54] 5. As far as at the real OS at the Cyprus–Rostov-on-Don path z-type disturbances are observed quite often(especially in the dawn-dusk periods of the day), one canconclude that they are generated by medium-scale TID withrelative amplitudes not less than 15–20%, this fact agreeingwell with the results presented by Stocker et al. [2000].

[55] As for the registration in the OS ionograms of severalhigh-angle rays in the form of a “comb” (see Figure 3a), wethink that such effect can be caused by a quasi-regular strati-fication of the electron concentration in the vicinity of the F -layer maximum with a vertical scale of ∼10–20 km. In thatcase, in the vicinity of the F -layer maximum, a comb-likestructure is formed with several local maximums of the elec-tron concentration, high-angle rays propagating along thesemaximums. One can assume that the diffuse background ac-companying the comb of the high-angle rays is determinedby the rays propagating in a combined way, that is part ofthe way the high-angle ray propagates along one ridge of theelectron concentration and then transits (for example, dueto a scattering) to another ridge etc. These assumptions areof a preliminary discussion character. Additional studies areneeded for a more substantiated conclusion.

[56] Thus, according to the results of continuous observa-tions during several months of the conditions of ionosphericHF propagation at the oblique sounding paths of variouslength and orientation, it is found that wave disturbancesare present almost permanently. The typical periods of wavedisturbances lie within the interval from 15 min to 1.5 hours.

[57] The variety of physical mechanisms of wave distur-bance formation of both natural and artificial origin (in-

14 of 18

Page 15: Oblique chirp sounding and modeling of ionospheric HF ...

GI2002 vertogradov et al.: modeling of ionospheric hf channel GI2002

Figure 8. Results of the modeling of ionograms of the oblique LFM sounding at the Cyprus–Rostov-on-Don path under the following parameters of TID: (a–c) δN = 20%, β = −30◦, T = 900 s, andΛ = 150 km; (d–f) δN = 20%, β = −60◦, T = 900 s, and Λ = 150 km. The duration of the imitationsession of the sounding was 300 s, the rate of the frequency sweeping was 100 kHz s−1. Within eachthree ionograms (Figure 8, a–c) and (Figure 8, d–f), the interval between the ionograms is 300 s.

15 of 18

Page 16: Oblique chirp sounding and modeling of ionospheric HF ...

GI2002 vertogradov et al.: modeling of ionospheric hf channel GI2002

Figure 9. Results of the modeling of ionograms of the oblique LFM sounding at the Cyprus–Rostov-on-Don path under the following parameters of TID: (a–c) δN = 15%, β = −60◦, T = 900 s, andΛ = 150 km; (d–f) δN = 20%, β = −60◦, T = 900 s, and Λ = 100 km. The duration of the imitationsession of the sounding was 300 s, the rate of the frequency sweeping was 100 kHz s−1. Within eachthree ionograms (Figure 9, a–c) and (Figure 9, d–f), the interval between the ionograms is 300 s.

16 of 18

Page 17: Oblique chirp sounding and modeling of ionospheric HF ...

GI2002 vertogradov et al.: modeling of ionospheric hf channel GI2002

cluding passage of the terminator, magnetic storms, vol-cano eruptions, hurricanes, thunderstorm activity, explo-sions and other impacts on the near-Earth medium) makesthe picture of wave-like disturbances appearance compli-cated and nonstationary. It is evident that this appearance iscaused by the presence of various relations in the solar wind-magnetosphere-ionosphere-atmosphere-Earth system. Thephysical basis of these relations is provided by generation,propagation, and dissipation of atmospheric waves of variousscales including planetary, tidal, acoustic-gravity, and infra-sonic waves. The dissipation of the energy of these wavesdetermines the essential part of the energetic balance of theupper atmosphere. At the presence in such complicated sys-tem of nonlinear interactions, an important role could beplayed by the trigger effect of a release of the accumulatedenergy, when local sources of a small power are able to causeconsiderable large-scale disturbances. For example, accord-ing to Chernogor [2003] in some cases the coefficient describ-ing the trigger effect of the energy release can reach valuesof 105–1010.

[58] In conclusion, we note that study of wave disturbancesis far for completing and requires systematic observationswith attraction of various methods at the boundary of geo-physics, physics of solar-terrestrial relations, seismology, andradio physics. In this aspect the results of systematic obser-vations of HF propagation conditions at the network of pathsof the oblique LFM sounding present obvious scientific andpractical interest and can be applied to studies of fine effectsrelated to the stratification of the ionosphere under impactof various types of wave processes caused by geomagneticand seismic activity, as well as by other phenomena.

[59] Later on we plan to perform the detailed analysis ofseparate events when the wave disturbances manifest them-selves in ionograms of oblique sounding. These manifesta-tions should be compared with the geophysical and seismicdata, the information concerning the rocket launches, andsome other peculiarities of the performed experiments. Theaim of such analysis is to try to single out on the backgroundof the dominated sources (like terminator and geophysicaldisturbances) some other sources and to estimate their in-put to the observed effects.

5. Conclusions

[60] Almost during the entire 2005, we conducted system-atic observations of the conditions of ionospheric propaga-tion of HF radio waves at the paths of various length andorientation. The results may be formulated in the followingway.

[61] 1. On the basis of the analysis of experimental andcalculated data of the oblique sounding at midlatitude pathsof various length and orientation, it is demonstrated that thebest results on forecasting of MUF of particular propagationmodes are obtained at correction of the IRI 2001 model bythe effective solar activity index.

[62] 2. The standard mean deviations of MOF fromtheir monthly mean mathematical expectations grow withan increase of the path length. Typical values of the

standard deviation values in quiet geophysical conditionsare estimated for paths of various length and orienta-tion. In the daytime the typical standard deviations ofMOF are: 1.2 MHz (Cyprus–Rostov-on-Don), 1.7 MHz(Inskip–Rostov-on-Don), 2.0 MHz (Norilsk–Rostov-on-Don), 2.2 MHz (Irkutsk–Rostov-on-Don), and 2.5 MHz(Magadan–Rostov-on-Don). At night the typical stan-dard deviations of MOF are 1.0 MHz (Cyprus–Rostov-on-Don), 1.5 MHz (Inskip–Rostov-on-Don), 1.5 MHz (Norilsk–Rostov-on-Don), 1.5 MHz (Irkutsk–Rostov-on-Don), and1.6 MHz (Magadan–Rostov-on-Don). No seasonal depen-dence of the standard deviation was found as a result of themeasurements at the Cyprus–Rostov-on-Don and Inskip–Rostov-on-Don paths.

[63] 3. At all paths almost always, MOF undergo short-period variations with quasi-periods from 20 min to 1.5 hour.The fluctuation amplitude can reach 2 MHz in quiet iono-spheric conditions at noon hours of the day. In separatedays in dawn-dusk hours at the passage of the terminator,the fluctuations at the Cyprus–Rostov-on-Don path can in-crease up to 5–8 MHz.

[64] 4. For various seasons, the spectra of MOF fluctua-tions are of a well-pronounced linear structure. The powerof spectral components of MOF fluctuations is concentratedin the 20–90 min range. The spectral composition of theMOF fluctuations varies from one day to another, thoughsome quasi-harmonic components can be present in MOFfluctuations spectra during several days.

[65] 5. It is found that TID with periods 15–30 min cor-relate well with sunrise and sunset for the middle point ofthe Cyprus–Rostov-on-Don path. The latter fact manifestsan important role of the mechanism of generation by theterminator of acoustic-gravity waves (AGW). TIDs are theionospheric response of AGW.

[66] 6. The imitation modeling confirmed that the z-typefeatures in ionograms are related to the traveling ionosphericirregularities. The modeling showed that the conditions forz-type features formation are rather tough and depend onthe relative amplitude of TID, wavelength of the distur-bance, and the direction of its propagation relative to theradio path orientation. On the basis of the comparisonof the experimental and calculated data of oblique sound-ing, TID parameters are estimated. It is shown that TIDwith the relative amplitude of electron density disturbanceδN ∼ 15−20%, wavelength of 150 km, and period of 15 minpropagating downward at the angle of 30◦–60◦to the hori-zon can be responsible for the observed z-type features at thefrequency-time ionogram registered at the Cyprus–Rostov-on-Don path.

[67] 7. The behavior of the measured frequency-amplitudeionograms can explain by the interference of nonseparateddiscrete rays, their phase not being a random value.

[68] Acknowledgment. The paper was supported by the

Russian Foundation for Basic Research, projects 05-05-08011 and

02-06-16075.

17 of 18

Page 18: Oblique chirp sounding and modeling of ionospheric HF ...

GI2002 vertogradov et al.: modeling of ionospheric hf channel GI2002

References

Al’pert, Ya. L. (1972), Propagation of Electromagnetic Wavesand Ionosphere (in Russian), 564 pp., Nauka, Moscow.

Barabashov, B. G., and G. G. Vertogradov (1996), Dynamicadaptive structural–physical model of decametre communica-tion channel, Math. Model (in Russian), 2, 3.

Bilitza, D. (2001), International reference ionosphere 2000, RadioSci., 36, 261.

Bilitza, D. (2002), Ionospheric models for radio propagationstudies, in The Review of Radio Science 1999-2002, edited byW. R. Stone, p. 625, IEEE Press, Piscataway, N. J..

Blagoveshchensky, D. V., and G. A. Zherebtsov (1987), High-Latitude Geophysical Phenomena and Forecasting of HF Chan-nels (in Russian), 272 pp., Nauka, Moscow.

Bowman, G. G. (1990), A review of some recent work on midlat-itude spread F occurrence as detected by ionosondes, J. Geo-magn. Geoelectr., 42, 109.

Burmaka, V. P., L. S. Kostrov, and L. F. Chernogor (2003),Statistical characteristics of signals of Doppler HF radar atsounding the middle ionosphere disturbed by starts of rocketsand solar terminator, Radiophys. Radio Astron (in Russian),8(2), 143.

Burmaka, V. P., V. I. Taran, and L. F. Chernogor (2004), Wave-like disturbances in the ionosphere, accompanying starts ofrockets on a background of natural transients, Geomagn. Aeron(in Russian), 44(4), 518.

Cherkashin, Yu. N., I. B. Egorov, V. P. Uryadov, andA. A. Ponyatov (2003), Experimental research of vari-ations of the maximum usable frequency the oblique soundingpaths, Radiophysics (in Russian), 46(12), 1011.

Chernogor, L. F. (2003), Physics of the Earth, atmosphereand geospace in view of a system paradigm, Radiophys. RadioAstron (in Russian), 8(1), 59.

Danilov, A. D., E. S. Kazimirovsky, G. V. Vergasova, andG. Ya. Khachikyan (1987), Meteorological Effects inthe Ionosphere (in Russian), 272 pp., Gidrometeoizdat, St.Petersburg.

Davies, K. (1969), Ionospheric Radio Waves, 502 pp., Blais-dell, Walham, Mass.

Erukhimov, L. M., et al. (1997), Pedersen mode ductingin randomly-stratified ionosphere, Waves Random Media, 7(4),531, doi:10.1088/0959-7174/7/4/002.

Filipp, N. D., N. Sh. Blaunshtein, L. M. Erukhimov,V. A. Ivanov, and V. P. Uryadov (1991), ModernMethods of Studies of Dynamical Processes in the Ionosphere(in Russian), 288 pp., Shtiintsa, Chisinau, Moldova.

Fukao, S., et al. (1991), Turbulent upwelling of the mid-latitude ionosphere: 1. Observation results by the MU radar,J. Geophys. Res., 96, 3725.

Gershman, B. N., L. M. Erukhimov, and Yu.Ya. Yashin (1984),The Wave Phenomena in the Ionosphere and Space Plasma (inRussian), 392 pp., Nauka, Moscow.

Goodman, J. M. (1992), HF Communication, Sci. and Tech-nol., New York.

Hearn, A. L., and K. C. Yeh (1978), A study of electron densityspectra of traveling ionospheric disturbances, J. Geophys. Res.,83(A4), 1442.

Ho, C. M., et al. (1998), Ionospheric total electron content

perturbations monitored by the GPS global network during twoNorthern Hemisphere winter storms, J. Geophys. Res., 103,26,409, doi:10.1029/98JA01237.

Hocke, K., and K. Schlegel (1996), A review of atmosphericgravity waves and traveling ionospheric disturbances: 1982–1995, Ann. Geophys., 14, 917.

Huang, C.-S., J. C. Foster, L. P. Goncharenko, G. J. Sofko,J. E. Borovsky, and F. J. Rich (2003), Midlatitude iono-spheric disturbances during magnetic storms and substorms, J.Geophys. Res., 108(A6), 1244, doi:10.1029/2002JA009608.

MacDougall, J. W., G. E. Hall, and K. Hayashi (1997), F -region gravity waves in the central polar cap, J. Geophys. Res.,102, 14,513, doi:10.1029/97JA01076.

MacDougall, J. W., D. A. Andre, G. J. Sofko, C. S. Huang,and A. V. Koustov (2001), Traveling ionospheric disturbanceproperties deduced from Super Dual Auroral Radar measure-ments, Ann. Geophys., 18, 1550, doi:10.1007/s00585-001-1550-z.

Moller, H. G. (1974), Backscatter results from Lindau. 1:The movement of curtains of intense irregularities in the polarF -layer, J. Atmos. Terr. Phys., 36(9), 1487, doi:10.1016/0021-9169(74)90227-X.

Poole, A. W. V., and G. P. Evans (1985), Advanced sounding:2. First results from an advanced chirp ionosonde, Radio Sci.,20, 1617.

Secan, J. A., and P. J. Wilkinson (1997), Statistical stud-ies of an effective sunspot number, Radio Sci., 32, 1717,doi:10.1029/97RS01350.

Shiokawa, K., C. Ihara, Y. Otsuka, and T. Ogawa (2003), Sta-tistical study of nighttime medium-scale traveling ionosphericdisturbances using midlatitude airglow images, J. Geophys.Res., 108(A1), 1052, doi:10.1029/2002JA009491.

Stocker, A. J., N. F. Arnold, and T. B. Jones (2000), Thesynthesis of traveling ionospheric disturbance (TID) signaturesin HF radar observations using ray tracing, Ann. Geophys., 18,54.

Taylor, M. J., J.-M. Jahn, S. Fukao, and A. Saito (1998), Pos-sible evidence of gravity wave coupling into the mid-latitude Fregion ionosphere during the SEEK campaign, Geophys. Res.Lett., 25, 1081, doi:10.1029/97GL03448.

Thomson, D. J. (1982), Spectrum estimation and harmonic anal-ysis, Proc. IEEE, 70, 1055.

Vertogradov, G. G. (2003), The simulator of broadbandionospheric radio channel, Radio Eng. Electron (in Russian),48(11), 1322.

Vertogradov, G. G. (2004), The simulator of decameter radiochannel, Radioelectronics (in Russian), 47(8), 51.

Vertogradov, G. G. (2005), Development of algorithms andsoftware of data processing of sounding of an ionosphere bysignals with linear-frequency modulation, scientific report (inRussian), 158 pp., Rostov State Univ., Rostov-on-Don, Russia.

Waldock, J. A., and T. B. Jones (1987), Source re-gions of medium scale traveling ionospheric disturbancesobserved at mid-latitudes, J. Atmos. Terr. Phys., 49(2),105, doi:10.1016/0021-9169(87)90044-4.

G. G. Vertogradov and V. G. Vertogradov, Rostov State Uni-versity, Rostov-on-Don, Russia.

V. P. Uryadov, Radiophysical Research Institute, Nizhny Nov-gorod, Russia.

18 of 18


Recommended