Rozprawy doktorskie na temat „Ionospheic wave propagation”

This thesis presents an update to the ionospheric propagation factor, M(3000)F2, global empirical model developed by Oyeyemi et al. (2007) (NNO). An additional aim of this research was to produce the updated model in a form that could be used within the International Reference Ionosphere (IRI) global model without adding to the complexity of the IRI. M(3000)F2 is the highest frequency at which a radio signal can be received over a distance of 3000 km after reflection in the ionosphere. The study employed the artificial neural network (ANN) technique using relevant geophysical input parameters which are known to influence the M(3000)F2 parameter. Ionosonde data from 135 ionospheric stations globally, including a number of equatorial stations, were available for this work. M(3000)F2 hourly values from 1976 to 2008, spanning all periods of low and high solar activity were used for model development and verification. A preliminary investigation was first carried out using a relatively small dataset to determine the appropriate input parameters for global M(3000)F2 parameter modelling. Inputs representing diurnal variation, seasonal variation, solar variation, modified dip latitude, longitude and latitude were found to be the optimum parameters for modelling the diurnal and seasonal variations of the M(3000)F2 parameter both on a temporal and spatial basis. The outcome of the preliminary study was applied to the overall dataset to develop a comprehensive ANN M(3000)F2 model which displays a remarkable improvement over the NNO model as well as the IRI version. The model shows 7.11% and 3.85% improvement over the NNO model as well as 13.04% and 10.05% over the IRI M(3000)F2 model, around high and low solar activity periods respectively. A comparison of the diurnal structure of the ANN and the IRI predicted values reveal that the ANN model is more effective in representing the diurnal structure of the M(3000)F2 values than the IRI M(3000)F2 model. The capability of the ANN model in reproducing the seasonal variation pattern of the M(3000)F2 values at 00h00UT, 06h00UT, 12h00UT, and l8h00UT more appropriately than the IRI version is illustrated in this work. A significant result obtained in this study is the ability of the ANN model in improving the post-sunset predicted values of the M(3000)F2 parameter which is known to be problematic to the IRI M(3000)F2 model in the low-latitude and the equatorial regions. The final M(3000)F2 model provides for an improved equatorial prediction and a simplified input space that allows for easy incorporation into the IRI model.

Statement of work: In essence the research work was to focus on the development of an ionospheric model suitable for real time HF frequency prediction and direction finding applications. The modelling of the ionosphere had to be generic in nature, sufficient to ensure that the CSIR could simultaneously secure commercial competitiveness in each of the three niche market areas aforementioned, while requiring only minimal changes to software architecture in the case of each application. A little research quickly showed that the development of an ionospheric model capable of driving a HFDFSSL system in "real time" would result in one having to make only slight re-structuring of the software to facilitate application of the same model in the areas of real time frequency prediction and spectrum management. The decision made at the outset of the project to slant the research toward the development of a model best suited for HF direction finding applications is reflected in the avenues followed during the course of the modelling process

Ionospheric phase delays of radio signals from Global Positioning System (GPS) satellites have been used to compute ionospheric Total Electron Content (TEC). An extended Chapman profle model is used to estimate the electron density profles and TEC. The Chapman profle that can be used to predict TEC over the mid-latitudes only applies during day time. To model night time TEC variability, a polynomial function is fitted to the night time peak electron density profles derived from the online International Reference Ionosphere (IRI) 2001. The observed and predicted TEC and its variability have been used to study ionospheric in°uence on Radio Astronomy in South Africa region. Di®erential phase delays of the radio signals from Radio Astronomy sources have been simulated using TEC. Using the simulated phase delays, the azimuth and declination o®sets of the radio sources have been estimated. Results indicate that, pointing errors of the order of miliarcseconds (mas) are likely if the ionospheric phase delays are not corrected for. These delays are not uniform and vary over a broad spectrum of timescales. This implies that fast frequency (referencing) switching, closure phases and fringe ¯tting schemes for ionospheric correction in astrometry are not the best option as they do not capture the real state of the ionosphere especially if the switching time is greater than the ionospheric TEC variability. However, advantage can be taken of the GPS satellite data available at intervals of a second from the GPS receiver network in South Africa to derive parameters which could be used to correct for the ionospheric delays. Furthermore GPS data can also be used to monitor the occurrence of scintillations, (which might corrupt radio signals) especially for the proposed, Square Kilometer Array (SKA) stations closer to the equatorial belt during magnetic storms and sub-storms. A 10 minute snapshot of GPS data recorded with the Hermanus [34:420 S, 19:220 E ] dual frequency receiver on 2003-04-11 did not show the occurrence of scintillations. This time scale is however too short and cannot be representative. Longer time scales; hours, days, seasons are needed to monitor the occurrence of scintillations.

We discuss various topics concerning the propagation, generation, and detec-tionof high-frequency (HF) radio waves in the Earth's ionosphere. With re-gardsto propagation, we derive a full wave Hamiltonian and a polarization evo-lutionequation for electromagnetic waves in a cold, stratified magnetoplasma.With regards to generation, we will be concerned with three experiments con-ducted at the ionosphere- radio wave interaction research facilities at Sura, Rus-siaand Tromsø, Norway. These facilities operate high power HF transmittersthat can inject large amplitude electromagnetic waves into the ionosphere andexcite numerous nonlinear processes. In an experiment conducted at the Surafacility, we were able to measure the full state of polarization of stimulatedelectromagnetic emissions for the first time. It is expected that by using thetechnique developed in this experiment it will be possible to study nonlinearpolarization effects on powerful HF pump waves in magnetoplasmas in the fu-ture.In another experiment conducted at the Sura facility, the pump frequencywas swept automatically allowing rapid, high-resolution measurements of SEEdependence on pump frequency with minimal variations in ionospheric condi-tions.At the Tromsø facility we discovered by chance a highly variable, pumpinduced, HF emission that most probably emanated from pump excited spo-radicE. Regarding detection, we have proposed a set of Stokes parametersgeneralized to three dimension space; and we have used these parameters in aninvention to detect the incoming direction of electromagnetic waves of multiplefrequencies from a single point measurement.

Radio signals transmitted by GPS satellites orbiting the Earth are modulated as they propagate through the electrically charged plasmasphere and ionosphere in the near-Earth space environment. Through a linear combination of GPS range and phase measurements observed on two carrier frequencies by terrestrial-based GPS receivers, the ionospheric total electron content (TEC) along oblique GPS signal paths may be quantified. Simultaneous observations of signals transmitted by multiple GPS satellites and observed from a network of South African dual frequency GPS receivers, constitute a spatially dense ionospheric measurement source over the region. A new methodology, based on an adjusted spherical harmonic (ASHA) expansion, was developed to estimate diurnal vertical TEC over the region using GPS observations over the region. The performance of the ASHA methodology to estimate diurnal TEC and satellite and receiver differential clock biases (DCBs) for a single GPS receiver was first tested with simulation data and subsequently applied to observed GPS data. The resulting diurnal TEC profiles estimated from GPS observations compared favourably to measurements from three South African ionosondes and two other GPS-based methodologies for 2006 solstice and equinox dates. The ASHA methodology was applied to calculating diurnal two-dimensional TEC maps from multiple receivers in the South African GPS network. The space physics application of the newly developed methodology was demonstrated by investigating the ionosphere’s behaviour during a severe geomagnetic storm and investigating the long-term ionospheric stability in support of the proposed Square Kilometre Array (SKA) radio astronomy project. The feasibility of employing the newly developed technique in an operational near real-time system for estimating and dissimenating TEC values over Southern Africa using observations from a regional GPS receiver network, was investigated.

Modelling ionospheric total electron content (TEC) is an important area of interest for radio wave propagation, geodesy, surveying, the understanding of space weather dynamics and error correction in relation to Global Navigation Satellite Systems (GNNS) applications. With the utilisation of improved ionosonde technology coupled with the use of GNSS, the response of technological systems due to changes in the ionosphere during both quiet and disturbed conditions can be historically inferred. TEC values are usually derived from GNSS measurements using mathematically intensive algorithms. However, the techniques used to estimate these TEC values depend heavily on the availability of near-real time GNSS data, and therefore, are sometimes unable to generate complete datasets. This thesis investigated possibilities for the modelling of TEC values derived from the South African Global Positioning System (GPS)receiver network using linear regression methods and artificial neural networks (NNs). GPS TEC values were derived using the Adjusted Spherical Harmonic Analysis (ASHA) algorithm. Considering TEC and the factors that influence its variability as “dependent and independent variables” respectively, the capabilities of linear regression methods and NNs for TEC modelling were first investigated using a small dataset from two GPS receiver stations. NN and regression models were separately developed and used to reproduce TEC fluctuations at different stations not included in the models’ development. For this purpose, TEC was modelled as a function of diurnal variation, seasonal variation, solar and magnetic activities. Comparative analysis showed that NN models provide predictions of GPS TEC that were an improvement on those predicted by the regression models developed. A separate study to empirically investigate the effects of solar wind on GPS TEC was carried out. Quantitative results indicated that solar wind does not have a significant influence on TEC variability. The final TEC simulation model developed makes use of the NN technique to find the relationship between historical TEC data variations and factors that are known to influence TEC variability (such as solar and magnetic activities, diurnal and seasonal variations and the geographical locations of the respective GPS stations) for the purposes of regional TEC modelling and mapping. The NN technique in conjunction with interpolation and extrapolation methods makes it possible to construct ionospheric TEC maps and to analyse the spatial and temporal TEC behaviour over Southern Africa. For independent validation, modelled TEC values were compared to ionosonde TEC and the International Reference Ionosphere (IRI) generated TEC values during both quiet and disturbed conditions. This thesis provides a comprehensive guide on the development of TEC models for predicting ionospheric variability over the South African region, and forms a significant contribution to ionospheric modelling efforts in Africa.

Thesis (Sc.D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1999.<br>Includes bibliographical references (p. 173-176).<br>by Michael James Starks.<br>Sc.D.

In ground based high frequency (HF) radio pumping experiments, absorption of ordinary (O) mode pump waves energises the ionospheric plasma, producing optical emissions and other effects. Pump-induced or natural kilometre-scale field-aligned density depletions are believed to play a role in self-focussing phenomena such as the magnetic zenith (MZ) effect, i.e., the increased plasma response observed in the direction of Earth's magnetic field. Using ray tracing, we study the propagation of ordinary (O) mode HF radio waves in an ionosphere modified by density depletions, with special attention to transmission through the radio window (RW), where O mode waves convert into the extraordinary (X, or Z) mode. The depletions are shown to shift the position of the RW, or to introduce RWs at new locations. In a simplified model neglecting absorption, we estimate the wave electric field strength perpendicular to the magnetic field at altitudes normally inaccessible. This field could excite upper hybrid waves on small scale density perturbations. We also show how transmission and focussing combine to give stronger fields in some directions, notably at angles close to the MZ, with possible implications for the MZ effect. In a separate study, we consider electromagnetic (e-m) beams with helical wavefronts (i.e., twisted beams), which are associated with orbital angular momentum (OAM). By applying geometrical optics to each plane wave component of a twisted nonparaxial e-m Bessel beam, we calculate analytically the shift of the beam's centre of gravity during propagation perpendicularly and obliquely to a weak refractive index gradient in an isotropic medium. In addition to the so-called Hall shifts expected from paraxial theory, the nonparaxial treatment reveals new shifts in both the transverse and lateral directions. In some situations, the new shifts should be significant also for nearly paraxial beams.

El sistema de comunicació ràdio d’alta freqüència (HF, en anglès) és usat arreu del món per agències governamentals i no governamentals sempre que calgui una alternativa a les comunicacions via satèl•lit: vaixells a alta mar, avions fora de cobertura de xarxes ràdio amb visió directa, operacions militars, zones on la infraestructura ha estat destruïda per algun tipus de desastre o bé zones llunyanes sense cap altre tipus de comunicació. La ràdio HF representa una alternativa, o un sistema de backup al satèl•lit per a comunicacions de llarg abast i en redueix els costos, evita la vulnerabilitat i els problemes de sobirania. En aquesta tesi s’ha estudiat l’enllaç HF entre la base antàrtica espanyola Juan Carlos I, situada a l’illa Livingston a l’arxipèlag de les Shetland del Sud, i Espanya. L’objectiu d’aquest treball és estudiar els problemes que afecten la propagació; és a dir, la relació senyal a soroll i interferència, la dispersió multicamí i la dispersió per efecte Doppler, i dissenyar la capa física d’un enllaç HF de baixa velocitat, poca potència i llarg abast. Pel que fa aquest últim punt es fan un parell de propostes: espectre eixamplat per seqüència directa (DSSS, en anglès) i multiplexació per divisió en freqüència ortogonal (OFDM, en anglès). El repte que es planteja és el de la definició de les característiques dels símbols que millor encaixen en aquest canal per tal d’obtenir un benefici de la diversitat temporal i freqüencial que ofereix el canal. Des de l’any 2003 diverses campanyes han permès estudiar aquest canal HF, però no va ser fins la campanya 2009/2010 que s’obtingué un foto de les característiques, diürnes i nocturnes, de la ionosfera. En els articles que es presenten en aquesta tesi hem estès el rang freqüencial d’estudi respecte a investigacions prèvies i hem mostrat diferències de comportament entre el dia i la nit. Hem usat els resultats de la caracterització del canal per a dissenyar i comparar la bondat dels símbols DSSS i OFDM. Ambdues possibilitats han resultat ser candidates a implementar l’enllaç HF entre l’Antàrtida i Espanya. Tot i així, ambdues tècniques representen visions diferents de la implementació del mòdem: mentre que DSSS obté bons resultats a baixa velocitat en entorns amb baixa relació senyal a soroll, OFDM aconsegueix tasses de velocitat més elevades en canals més benignes.<br>Los sistemas de radio de alta frecuencia (HF, en inglés) son usados por agencias gubernamentales y no gubernamentales en todo el mundo siempre que se necesite una alternativa a las comunicaciones por satélite: barcos en alta mar, aviones fuera del rango de cobertura de las redes radio de visión directa, operaciones militares, zonas donde la infraestructura ha sido destruida por algún desastre. Ésta ofrece una alternativa, o representa un sistema de backup, a las comunicaciones vía satélite, evitando los costes, la vulnerabilidad y los problemas de soberanía de las comunicaciones por satélite. En esta tesis se ha estudiado el enlace HF entre la base antártica española Juan Carlos I en la isla Livingston, en las Shetland del sur y España. El objetivo de este trabajo es el estudio de las limitaciones de la propagación ionosférica (como la relación señal a ruido e interferencia, la dispersión multicamino y la dispersión por efecto Doppler) y el diseño de la capa física de un enlace HF de baja velocidad, baja potencia y largo alcance. Se han estudiado un par de propuestas para este enlace, como son el espectro ensanchado por secuencia directa (DSSS, en inglés) y la multiplexación por división en frecuencia ortogonal (OFDM, en inglés). El reto ha sido definir las características que mejor se adecuan a este enlace para poder aprovechar la diversidad temporal y frecuencial que ofrece el canal HF. Desde el año 2003 diversas campañas de sondeo han permitido estudiar el canal HF pero no es hasta la campaña 2009/2010 que se consigue una fotografía de la actividad ionosférica tanto nocturna como diurna. En los artículos que se presentan en esta tesis hemos extendido los estudios previos a todo el rango de frecuencias HF y hemos mostrado las diferencias entre el día y la noche. Hemos usado estos resultados de caracterización del canal para diseñar y comparar símbolos DSSS y símbolos OFDM. Ambas posibilidades han resultado ser posibles candidatas para implementar un enlace HF de baja velocidad entre la Antártida y España. Sin embargo ambas técnicas representan dos aproximaciones distintas a la implementación del módem. Mientras que DSSS consigue un buen funcionamiento a baja velocidad en escenarios con baja relación señal a ruido, OFDM consigue tasas de transmisión más altas en escenarios más benignos.<br>High Frequency (HF) radio is used by governmental and non nongovernmental agencies worldwide whenever an alternative to satellites for sky wave communication is needed: ships at sea, aircraft out of range of line-of-sight radio networks, military operations, disaster areas with communication infrastructure destroyed or distant regions lacking other communications. It offers an alternative to satellites, or a backup, for long-haul communications, thus avoiding the costs, vulnerabilities and sovereignty concerns of satellite communications. In this thesis the HF link between the Antarctic Spanish Station Juan Carlos I in Livingston Island, South Shetlands and Spain is studied. The aim of this study is to address the impairments that affect HF propagation (i.e., signal-to-noise plus interference ratio, multipath and Doppler shift and spread) and to design the physical layer of a low rate, low power and long-haul HF link. Some proposals regarding this last issue are addressed, i.e., direct sequence spread spectrum (DSSS) and orthogonal frequency division multiplexing (OFDM). The challenge is to define the symbol characteristics that best fit the link to benefit from time and frequency diversity that offers the HF channel. Since 2003 several transmission campaigns have allowed to study the HF channel but it is not until the 2009/2010 campaign that we have achieved a whole picture of both diurnal and nocturnal ionospheric activity. In the papers presented in this thesis we have extended the previous research to the whole range of HF frequencies and we have shown the differences on performance between day and night. We have used the results from channel characterization to design and compare the performance of DSSS and OFDM symbols. Both techniques have turned out to be possible candidates to implement a low rate HF link between Antarctica and Spain. However, both techniques stand for different approaches of the modem: DSSS achieves good performance at low data rate in low SNR scenarios, whereas OFDM achieves higher data rate in benign channels

Une formulation de la densite electronique en fonction de l'altitude et de l'angle solaire est etablie pour la basse ionosphere, cette formulation est basee sur la resolution des equations differentielles relatives aux densites des charges caracteristiques du milieu ionise traite. La relation entre les problemes aeronomiques et les problemes de propagation des ondes electromagnetiques est etudiee en reliant la vitesse de phase des ondes tbf a certains parametres representatifs de l'atmosphere moyenne. Cette relation est ensuite utilisee pour evaluer les variations de l'altitude de reflexion des ondes tbf ainsi que pour la determination de la temperature dans la mesosphere

博士<br>國立中央大學<br>太空科學研究所<br>92<br>In this dissertation, a procedure deriving the vertical group and phase velocities of periodic waves in the ionosphere is proposed. The method is based on Fourier analysis and considered the difference of angular frequency and wave number as a function of number of data points used in a spectral analysis. Simulations and observations from ionosondes and a VHF radar are investigated to understand the vertical propagation and source of the waves. The simulation is employed to demonstrate the usefulness of the procedure while some noises are added and multiplied to the wave packet for examing the suitability and stability of the procedure. The developed procedure is further applied to analyze waves appearing in transient layers during a solar eclipse and appearance of an additional sub-layer as well as a permanent layer of the F-region. A study of the waves caused by repid change of sunlight during the solar eclipse shows that the energy source lies at the altitude near the F1 peak. An examination of the waves during appearance of the additional layer in the equatorial regions indicates the wave source is around the peak of the layer. An analysis of the data observed from ionosonde and the VHF radar, the F2 peak is an energy source of the associated waves. Results of this dissertation demostrate the waves sources are mainly around the peaks of layers in the ionosphere.

Very Low Frequency (VLF) radio waves propagate within the Earth-ionosphere waveguide with very little attenuation. Modifications of the waveguide geometry affect the propagation conditions, and hence, the amplitude and phase of VLF signals. Changes in the ionosphere, such as the presence of the D-region during the day, or the precipitation of energetic particles, are the main causes of this modification. Using narrowband receivers monitoring remote VLF transmitters, the amplitude and phase of these signals are recorded. A multivariate data analysis technique, Principal Component Analysis (PCA), is applied to the data in order to determine parameters such as seasonal and diurnal changes which affect the variation of these signals. Data was then analysed for effects from extragalactic gamma ray bursts, terrestrial gamma ray flashes and solar flares. Only X-rays from solar flares were shown to have an appreciable affect on ionospheric propagation.<br>Thesis (M.Sc.)-University of KwaZulu-Natal, Westville, 2009.

Wave induced electron precipitation (WIEP) can modify the ionosphere above a sub-ionospherically propagating VLF signal in such a way as to perturb the amplitude and phase of the signal: The "Trimpi Event". In this thesis trimpi events are used in a study of WIEP events and in the responsible mechanism: The gyroresonant interaction. Trimpi activity at middle latitudes (SANAE, Antarctica, L = 4.02) and low latitudes (Durban, RSA , L = 1.69) together with the corresponding theory for the gyroresonant interaction is examined and compared. A newly developed computerised system for the detection and analysis of trimpi events has been developed in Durban. This system has been used to analyse tape data recorded at SANAE. Trimpi events were found on various transmitter paths to SANAE and a complete study of 1982 data has led to the establishment of trimpi characteristics as seen at SANAE: an absence of positive events and causative whistlers, a preference for short duration events (t < 25s), the occurrence of some very large events (up to 90% signal attenuation) , two minima in occurrence near 0015 and 0400 h Local Time, low occurrence and occurrence rate of events and evidence that interactions with non-ducted whistlers are of importance. The computerised sytem was then extended to collect data at Durban simultaneously from up to 20 transmitters worldwide. Examination of data from this survey showed very low occurrence rates of trimpis but yielded some daytime events for which the effectiveness of the gyroresonance interaction, which successfully explains the trimpi event at middle and low latitudes, had to be questioned. Thus a fully relativisic test particle simulation of the gyroresonant interaction was used to examine the effectiveness of gyroresonance at low L for producing trimpi events. This simulation was run for a wide range of interaction parameters and yielded the following constraints for effective pitch angle scattering (and hence precipitation) of electrons at low L: wave intensities in excess of 150 nT, wave frequencies in excess of 10 kHz and background electron densities at least one order of magnitude higher than normal. First data from the OMSKI project, a sophisticated VLF receiver operated at Durban as part of an international project, shows further evidence of low-latitude trirmpi activity. A survey of one month's continuous data is presented. In face of the evidence that trimpi events that occur at low L have the same signature as those at middle L but that the standard gyroresonance interaction is insufficient to cause them, alternate scenarios that could enhance the interaction were sought. In particular distortions in the ambient magnetic field (eg. PC-5 pulsations) were modelled using a new dipole-like background field model. This simulation showed that distortions which tend to reduce magnetic field curvature along field lines can significantly enhance the gyroresonant conditions and hence the interaction. A new set of conditions for effective gyroresonance at low L is thus established and contrasted with the more lenient conditions at middle L. A study of "frequency tracking" as a means to prolong resonance showed that natural whistlers do not posess the required frequency /time characteristics for this mechanism, and that artificial waves in a narrow range around the equatorial resonance frequency would ~ well suited for this purpose. An overview of the status of worldwide Trimpi detection networks together with the S.P.R.I. 's role in this regard is presented.<br>Thesis (Ph.D.)-University of Natal, Durban, 1991.

Research Doctorate - Doctor of Philosophy (PhD)<br>Ultra low frequency (ULF) waves are generated by the interaction of the solar wind with Earth’s magnetosphere. The waves propagate through the ionosphere and may be seen in data from SuperDARN Over-The-Horizon radars as oscillations in Doppler velocity which coincide with oscillations of the geomagnetic field recorded by ground magnetometers. Data from the two Tasman International Geospace Environment Radars (TIGER) and the magnetometer located on Macquarie Island over 2006-2009 show that ULF wave signatures are common. The dependence with time of day showed most ULF radar events occurred between 6-12 UT (dusk and pre-mid-night sector). Using the Maximum Entropy Method (MEM) several spectral aspects of detected ULF waves were investigated for the first time. Most of the recorded frequencies were in the range 1- 4 mHz. Some frequencies were seen more often than others with the more common frequencies spaced about 0.3mHz apart, consistent with previous studies. There was no evidence of a frequency dependence on latitude. Field line resonance (FLR) signatures in the radar data were expected but were not common, as shown by comparisons with coincident ground magnetometer data. This study has shown that most radar echoes containing ULF wave Doppler signatures appear to be backscattered from the plasmapause region mapped into the ionosphere. Modelling of the Doppler velocity combined with a ULF wave model and two dimensional ray tracing was achieved for the first time. The results are consistent with experimental observations. With further improvement, this modelling should provide a valuable tool to understand shear Alfven mode and fast mode plasma wave interactions within the ionosphere and their effects on HF propagation through the ionosphere.

"July, 2003"<br>Includes bibliographical references (leaves 177-183)<br>xxii, 183 leaves : ill. (some col.) ; 30 cm.<br>Title page, contents and abstract only. The complete thesis in print form is available from the University Library.<br>Thesis (Ph.D.)--University of Adelaide, School of Chemistry and Physics, Discipline of Physics and Mathematical Physics, 2004

Nell'analisi di routine di segnali sismici registrati sia da reti fisse sia temporanee è abbastanza comune riconoscere, alle varie stazioni, coppie (doublets) o gruppi (multiplets) di eventi sismici caratterizzati da forme d'onda molto simili. Tali eventi se localizzati in corrispondenza di una ristretta fascia geografica costituiscono, se caratterizzati da comune meccanismo sorgente, quelle che vengono definite “famiglie sismogenetiche” (Tsujiura, 1983). Il riconoscimento di terremoti caratterizzati (se registrati ad una comune stazione) da sismogrammi molto simili consente indagini di dettaglio in grado di fornire informazioni molto più accurate rispetto ad una semplice localizzazione di routine; determinare una relazione diretta tra famiglie sismogenetiche e strutture geologiche presenti in una determinata area consente infatti di discriminare all’interno della stessa la presenza di sorgenti potenzialmente attive. In particolare, uno studio prolungato nel tempo dell’occorrenza di terremoti aventi le caratteristiche sopra riportate rappresenta lo strumento fondamentale al fine della determinazione dei periodi di riattivazione delle sorgenti individuate. L’analisi di somiglianza di forme d’onda può essere implementata sia nel dominio del tempo sia nel dominio delle frequenze, utilizzando a seconda dei casi tecniche basate rispettivamente sulle funzioni di cross correlazione e di cross spettro. In caso di analisi nel dominio delle frequenze i segnali analizzati sono ottenuti a partire da sismogrammi convertiti in spettri tramite l’utilizzo della funzione trasformata di Fourier. A causa del non sempre soddisfacente rapporto segnale disturbo caratterizzante le registrazioni fornite dalle reti sismiche fisse RSNI (Rete Sismica Italia Nord Occidentale) ed RSLG (Rete sismica Lunigiana-Garfagnana), gestite direttamente dalla sezione geofisica dell’Università di Genova ed utilizzate come base di partenza per gli studi effettuati, nel presente lavoro di tesi sono state implementate tecniche di analisi del segnale sismico esclusivamente nel dominio del tempo. Al fine di determinare in termini oggettivi un adeguato settaggio per tutti i parametri coinvolti in un’analisi di somiglianza di forme d’onda si è compiuto uno studio di dettaglio considerando come test la sismicità dell’Appennino Settentrionale (area Lunigiana- Garfagnana) registrata dalla rete fissa RSLG nel periodo 1999 - 2003. In base alle risultanze derivate da una preliminare analisi del rapporto segnale disturbo effettuata a ciascuna stazione della rete RSLG, è stato selezionato un data set composto da circa 1.000 terremoti, aventi magnitudo locale compresa tra 1.5 e 4.1. L’analisi delle forme d'onda è stata implementata nel dominio del tempo utilizzando la funzione di cross correlazione normalizzata. A differenza di recenti studi su doublets sismici, per ogni singolo sismogramma è stata analizzata una finestra temporale comprendente tutte le fasi del segnale considerato. Molteplici test eseguiti hanno infatti dimostrato come effettuare un’analisi di somiglianza considerando ridotte porzioni di sismogramma (pochi secondi) conduca in modo inequivocabile alla determinazione di famiglie sismogenetiche erroneamente sovradimensionate. Come sottolineato in Ferretti et al., 2005 (accettato in via preliminare per la pubblicazione sulla rivista “Bulletin Seismological Society of America”) l’indice di cross-correlazione calcolato considerando un’adeguata finestra temporale, comprendente fase P, fase S e parte della coda, consente di ottenere valori di similitudine dipendenti anche da componenti di segnale strettamente legate alla propagazione nel mezzo. Una volta ottenuti affidabili indici di somiglianza si è proceduto alla determinazione, a partire da coppie di terremoti simili (doublets), di una soglia minima di cross correlazione (indice di somiglianza) riferita a ciascuna stazione utilizzata, da applicare al fine di discriminare quelle che vengono definite famiglie sismogenetiche (multiplets). A tale scopo è stata applicata una innovativa procedura (Ferretti et al., 2005) in grado di considerare sia le componenti verticali sia le componenti orizzontali dei segnali registrati a tutte le stazioni della rete sismica considerata. Il raggruppamento in famiglie è stato effettuato utilizzando la “bridging technique” (Aster and Scott, 1993), considerandone vantaggi e svantaggi rispetto alle metodologie classiche. Il risultato finale è stata la determinazione di 27 multiplets, ognuno dei quali riconosciuto da più stazioni con valori minimi di somiglianza superiori all’ 80%. Considerando le famiglie sismogenetiche maggiormente significative (in termini di componenti) e contemporaneamente localizzate internamente alla rete RSLG, è stata applicata a ciascuna di esse una procedura di localizzazione in relativo tramite l’utilizzo del “double difference algorithm” (Waldhauser and Ellsworth, 2000). L’accuratezza dei parametri ipocentrali ottenuta tramite la procedura di ri-localizzazione è stata testata analizzando due differenti data set, composti rispettivamente da eventi localizzati internamente ed esternamente alla rete RSLG. I risultati ottenuti hanno messo in evidenza la buona affidabilità del metodo per eventi sismici localizzati internamente al network ed allo stesso tempo una scarsa significatività dei risultati considerando un data set caratterizzato al contrario da eventi sismici caratterizzati da elevato gap azimutale e non trascurabile distanza ipocentro-prima stazione. Per ciascuna famiglia rilocalizzata è stato calcolato, ove possibile, il meccanismo focale cumulato al fine di determinare l’orientazione del piano di faglia principale. Una successiva applicazione delle metodologie sopra descritte è stata effettuata utilizzando come data set di partenza circa 250 terremoti, registrati dalla rete sismica RSNI nel periodo Agosto 2000 - Luglio 2001, localizzati in una ristretta area geografica ubicata pochi km a NO di Acqui Terme (Monferrato, Piemonte) (Massa et al., 2005, accettato in via preliminare per la pubblicazione sulla rivista “Journal of Seismology”). L’analisi di doublets unita ad una successiva procedura di localizzazione in relativo ha condotto alla determinazione di 5 multiplets, ognuno dei quali aventi parametri di localizzazione caratterizzati, se paragonati ai medesimi derivati dalla localizzazione di bollettino, da un brusco decremento degli errori. Le nuove localizzazioni delle famiglie sismogenetiche, nonostante abbiano consentito di definire per l’area in studio una distribuzione di sismicità interpretabile in riferimento alle conoscenze concernenti l’assetto geologico strutturale dell’area stessa, rimangono in questo caso affette, per quanto riguarda le coordinate assolute, da non trascurabili errori insiti nelle localizzazioni di partenza. La configurazione del network a disposizione, sia in termini di gap azimutale sia in termini di distanza epicentri-stazione, non ha consentito, relativamente alla posizione dell’area epicentrale, di ottenere affidabili localizzazioni assolute. Il confronto dei risultati derivanti dall’applicazione di medesime metodologie, a partire da differenti condizioni al contorno, ha consentito di definire le condizioni limite di applicabilità delle stesse, le quali se utilizzate senza alcun criterio di selezione condurranno in generale ad un mancato miglioramento delle condizioni di partenza e/o a risultati erronei. La parte conclusiva degli studi trattati è stata dedicata allo sviluppo di un nuovo algoritmo di localizzazione assoluta basato esclusivamente su un’analisi di somiglianza di forme d’onda effettuata ad una singola stazione di riferimento (Massa et. al. 2005, sottomesso alla rivista “Journal of Geophysical Research”). Tale procedura è stata implementata considerando come riferimento la stazione mono-componente di Sant’Anna di Valdieri (rete sismica RSNI), ubicata nelle Alpi Sud Occidentali, in prossimità del confine italo-francese. Sono stati raccolti in un data base di partenza, accuratamente selezionato a seguito di un’analisi del rapporto segnale disturbo, circa 2.700 sismogrammi verticali, registrati in un’area di 40 km x 40 km nel periodo 1985-2004. L’analisi di somiglianza, precedentemente descritta è stata in grado di discriminare per il periodo considerato 80 multiplets, a ciascuno dei quali è stato possibile associare un evento master (evento di riferimento) in corrispondenza del quale fare collassare le coordinate ipocentrali di tutti gli eventi appartenenti alla famiglia associata al medesimo. Utilizzando il data set di multiplets ricavato per il periodo in esame, l’algoritmo di localizzazione è stato testato utilizzando un data set ridotto composto da circa 100 terremoti, registrati nell’area in studio nel periodo Gennaio 2003 - Giugno 2004. Tramite la suddetta metodologia di localizzazione, basata esclusivamente sui risultati derivati da un’analisi di forme d’onda nel dominio del tempo, è stato possibile localizzare circa il 50% degli eventi appartenenti al data set ridotto. Il vantaggio principale di tale procedura, rispetto alle tecniche usualmente utilizzate per determinare le coordinate ipocentrali di eventi sismici, risulta l’indipendenza della medesima da errori derivanti da sfavorevoli geometre di rete rispetto all’area epicentrale, dal numero di stazioni (fasi registrate), dalla distanza ipocentro-prima stazione e da errori di lettura delle fasi sismiche da parte di un operatore. L’attenzione rivolta allo studio delle sequenze sismiche, con la conseguente possibilità di dare una corretta caratterizzazione alle strutture sismogenetiche presenti nelle aree considerate, rappresenta uno passo fondamentale per qualsiasi tipo di studio disciplinare successivo; la qualità dei risultati ottenibili attraverso la costruzione di modelli tomografici e di propagazione (sulla base dei quali vengono effettuati gli studi di rischio sismico) dipende infatti dagli errori ottenuti durante le procedure di localizzazione dei terremoti e di conseguenza dalla corretta individuazione delle strutture sismogenetiche responsabili degli stessi.<br>Istituto Nazionale di Geofisica e Vulcanologia, sezione di Milano-Pavia<br>Unpublished<br>3.1. Fisica dei terremoti<br>open