© K. Paul et al., Published by EDP Sciences 2025

1 Introduction

Ionospheric soundings often detect F trace spreading echoes in ionograms, referred to as spread F (SF) (e.g., Berkner & Wells, 1934; Booker & Wells, 1938; Bowman, 1981). SF-related plasma irregularities can degrade modern radio technologies, satellite navigation systems, and the reliability of remote-sensing data (e.g., Vasylyev et al., 2022), highlighting the importance of understanding their occurrence, drivers, and long-term variability. Different types of SF observed in ionograms can be attributed to multiple structured diffused patterns pointing to plasma irregularities, which may last from a few minutes to several hours (e.g., Herman, 1966; Bowman, 1991, 1998; Hajkowicz, 2007; Lynn et al., 2011; Paul et al., 2019). SF phenomena are usually classified into two types: Range SF (RSF), observed around the lower frequency band of the ionogram F-region traces and representing diffused echoes around 10 km or more over the true range of the main echo; and Frequency SF (FSF), observed as diffused echoes near the critical frequency of the F trace, with less spreading in the lower frequency band of the ionogram. Bowman (1960) suggested that RSF is associated with the frontal structure of irregularities linearly extending around 1,000 km or more, while FSF is attributed to diffused echoes scattered from the ionosphere close to the zenith. King (1970) observed SF development over several mid-latitude stations in Australia and New Zealand, attributing RSF to ridges of F-region ionization, formed by large-scale tilts of the bottom-side F-layer isodensity contours, and concluded that FSF is the decay product of RSF.

Mid-latitude SF has been studied since the 1960s over different latitude and longitude sectors (e.g., Singleton, 1968; King, 1970; Bowman, 1990, 1996; Paul et al., 2018, 2019, 2022, 2023). These studies highlight the main morphological features of mid-latitude SF, including seasonal and diurnal characteristics, and nighttime ionospheric background conditions favorable for SF generation (Huang et al., 2011), and underline the complexity of the SF triggering mechanisms (Fejer & Kelley, 1980; Hajkowicz, 2007). Long-term analyses indicate anticorrelation of SF occurrence with solar flux, although this relationship exhibits latitude-dependent characteristics (Bowman, 1996). This latitudinal variation is evident in the European sector, where Paul et al. (2018) showed that SF activity differs between mid-high and mid-low latitudes. Further supporting this, a subsequent study by Paul et al. (2022) on the RSF and FSF occurrence over the European mid-latitudes provides evidence that SF events are governed by different underlying mechanisms at higher versus lower mid-latitudes.

Plasma irregularities causing mid-latitude SF may originate from gravity wave (GW) propagation (Miller, 1997; Miller et al., 1997; Bhaneja et al., 2009), with the Perkins instability contributing to their formation (Perkins, 1973), particularly at nighttime. GWs are considered an important trigger of SF (Baker & Davies, 1969; Wan et al., 1998; Belashov & Belashova, 2015), especially at mid-latitudes (Huang et al., 1994), where the Perkins instability growth rate is small. Travelling Ionospheric Disturbances (TIDs) are the ionospheric manifestation of internal Atmospheric Gravity Waves (AGWs) in the neutral atmosphere (e.g., Hines, 1963; Dyson et al., 1970), and their activity, particularly medium-scale TIDs (MSTIDs), is an established SF driver. Large- and medium-scale ionospheric irregularities are the dominant structures generating mid-latitude SF traces, and it has been shown that TIDs wave trains precede SF events (Bowman, 1990). Studies of the longitudinal variation of the vertical Total Electron Content (vTEC) estimated by the Global Navigational Satellite System (GNSS) show that MSTID signatures precede the onset of SF (Kotake et al., 2006). Similarities in SF activity at nearby European mid-latitude ionosonde stations, affected by common TID activity, support the hypothesis that MSTIDs drive the SF development (Paul et al., 2023). Another indicator for evaluating the dynamics and spatial scale of ionospheric irregularities is the Rate of TEC Index (ROTI). This GNSS-based metric quantifies the severity of amplitude and phase fluctuations in GNSS signals, and directly correlates with radio wave scintillation (Pi et al., 1997). ROTI has been extensively used to characterize ionospheric irregularities mainly at equatorial- and high-latitudes (Tiwari et al., 2013; Cherniak et al., 2014, 2018; Kotulak et al., 2020). However, fewer studies have applied it at mid-latitudes due to lower frequency and impact of scintillation (Vadakke Veettil et al., 2017). ROTI can also be used as a potential indicator of SF, as European ionosonde observations show RSF coinciding with significant ROTI values (>0.15 TECU/min) and MSTID activity (Paul et al., 2024).

Previous studies have shown that mid-latitude SF occurrence anticorrelates with solar flux (Bowman, 1996), a trend also evident in long-term variations of MSTID activity driven by the Perkins instability. Nighttime MSTID activity at mid-latitudes increases during periods of low solar activity, when electrodynamical forces such as the Perkins instability become more effective (Otsuka et al., 2021). This behavior is consistent with the influence of the Upper Atmospheric Neutral Particle Density (UA-NPD) on TID amplitudes (Hines, 1963), since secondary GWs generated by the dissipation of primary GW propagating from below increase under low solar activity (Otsuka et al., 2021).

The aim of this research is to evaluate in detail the climatology of nighttime SF observed at the mid-latitude ionospheric station in Roquetes, Spain. The analysis is based on a unique long-term ionogram dataset covering 1955–2022, complemented by GNSS-derived detrended TEC (d-TEC) and ROTI data for 2012–2016. We investigate the diurnal, seasonal, and Solar-Cycle variability of both Range Spread F (RSF) and Frequency Spread F (FSF), their dependence on solar activity, and the seasonal patterns of MSTID-driven SF. Additionally, we quantify the coincidence of SF and MSTID events and explore the potential relationship between significant ROTI values and SF activity. This comprehensive approach allows us to assess the dominant electrodynamic drivers of mid-latitude SF, clarify the links between SF and MSTID activity, and provide new constraints for understanding ionospheric dynamics.

This manuscript is organized as follows: Section 2 describes the dataset and methodology. Section 3 presents the analysis results and the implications of our findings, including diurnal and seasonal patterns of MSTID-driven SF occurrence and its dependence on solar activity, as well as the potential link of significant ROTI values with SF occurrence. Section 4 provides concluding remarks and summarizes the key contributions of this research.

2 Data and method

Ionograms recorded by the ionosonde of the Ebro Observatori (EO) in Roquetes, Spain (40.8°N, 0.5°E geographic; geomagnetic latitude: 42.69°N) were analyzed to identify SF activity. Furthermore, d-TEC and ROTI maps over Europe, obtained from the Dense Regional and Worldwide International GNSS-TEC observation, the DRAWING-TEC¹ project (Tsugawa et al., 2018) at a 10-min resolution, were employed to identify MSTIDs and significant ROTI values. Solar activity has been considered in this study by using the monthly and yearly average sunspot numbers (SSN), obtained from the Sunspot Index and Long-term Solar Observations (SILSO²), provided by the Royal Observatory of Belgium (Clette & Lefèvre, 2015). Additionally, we considered geomagnetic activity conditions based on the hourly average of the Disturbance storm time index (Dst), obtained from the Geomagnetic Data Service of the World Data Center for Geomagnetism in Kyoto³, Japan.

2.1 Ionosonde data

In this study, we used long-term ionosonde data from the EO, which began ionospheric studies and observations just in time for the International Geophysical Year (1957–1958). Beginning at that time, a new ionospheric section installed the first vertical ionospheric sounder of the Iberian Peninsula in 1955 (Batlló & Altadill, 2007). The ionosonde received the URSI code EB040, and the first ionograms were recorded in March 1955 by the French sounder ST35. The instrument operated in the frequency range of 1.4–16 MHz and is described in Cardús (1955). This ST35 sounder was replaced by a Swedish ionosonde from “Magnetic AB” in 1967, which sounded in the frequency range of 0.25–20 MHz and remained in operation until 1987. An agreement with the University of Massachusetts Lowell (USA) made it possible to install a Digisonde 256 (DGS 256) in 1988, which ran until 2011. The DGS 256 had an adjustable sounding frequency range from 1 to 30 MHz. A DPS-4D⁴ ionosonde is currently operating since 2011. A catalogue⁵ of all ionospheric measurements recorded since 1955 and a data repository is maintained at the EO and is available for ionospheric studies. The EB040 ionospheric station has contributed hourly measurements to the Ionospheric Digital Database of Worldwide Vertical Incidence Parameters⁶ of the National Centers for Environmental Information (NCEI), formerly the National Geophysical Data Center (NGDC) of the National Oceanic and Atmospheric Administration (NOAA). According to this dataset, FSF occurrence is obtained and identified from the measured values of the critical frequency of the F2 layer (foF2) which are qualified with the letter F (Piggott & Rawer, 1972) after ionogram scalings done by trained experts. This FSF occurrence data encompasses the years 1955–2019 and is analyzed for long-term and solar activity variation. Furthermore, digital ionograms recorded at EB040 from 1988 to the present are available through the Digital Ionogram Data Base (DIDBASE⁷) of the Global Ionospheric Radio Observatory (GIRO⁸) portal. A detailed investigation of nighttime SF occurrence (from dusk to dawn), including both RSF and FSF types, was carried out for the period 2012–2022. This ionogram dataset spans eleven years, equivalent to one Solar Cycle, and all ionograms were recorded using the same instrument (DPS-4D), operating at a 5-min cadence. For this study, every single ionogram recorded during this interval – about one million in total – was examined to identify FSF and RSF events. Only cases in which SF persisted for at least 15 min, i.e., three or more consecutive ionograms with 5-min sampling, were considered for analysis.

2.2 GNSS maps data

In this study, we also used d-TEC and ROTI data as indicators of MSTID activity and ionospheric irregularities. We did not calculate the d-TEC and ROTI values; instead, we used the d-TEC⁹ and ROTI¹⁰ maps provided by the DRAWING-TEC project (Tsugawa et al., 2018). d-TEC and ROTI are obtained from the calculated TEC values. The phase differences obtained from dual GNSS receivers between a given satellite-receiver pair allow the TEC to be obtained in the line of sight, also known as slant TEC, sTEC (e.g., Wanninger, 1993). Although the process of calculating d-TEC and ROTI maps is not the subject of this research, we summarize it briefly below and refer the reader to the literature provided for further information.

The method to obtain the d-TEC maps is very well described by Otsuka et al. (2013) and Otsuka et al. (2021). In short, to extract the perturbation component, the sTEC time series for each satellite-receiver pair is detrended by subtracting a 1-h running average (centered ±30 min around the epoch of interest). This step isolates the shorter-period fluctuations likely associated with MSTIDs, while filtering out longer-term trends and background ionospheric variations. The resulting sTEC values are then converted to vertical vTEC by applying a slant factor and are referenced to the point where the satellite receiver’s line of sight intersects with the reference shell at an altitude of 300 km. This point is known as the ionospheric pierce point (IPP). The data with satellite elevation angles below 35° are excluded in this procedure to ensure mapping accuracy. All TEC values from multiple satellites and receivers within each 0.15° × 0.15° pixel are smoothed with a 10-min running average to address the uneven distribution of IPPs and to compensate for the scarcity of the TEC data distribution. Finally, anisotropic smoothing is also applied to adapt the meridional convergence of the grid spacing and ensure consistent spatial resolution. In this way, d-TEC values are mapped on an ionospheric shell at 300 km altitude, with a pixel size of 0.15° × 0.15° in latitude and longitude, using measurements of more than 800 GNSS receivers distributed across Europe (Otsuka et al., 2011, 2013), providing maps with an approximate spatial resolution of 80 × 80 km and a temporal resolution of 10 min. According to Otsuka et al. (2013), the precision of the TEC disturbance detected in d-TEC maps is, in theory, on the order of 0.01–0.02 TECU (1 TECU = 10¹⁶ electrons/m²).

d-TEC values represent fluctuations in the TEC relative to its background level (Otsuka et al., 2013) and are widely used as an indicator of MSTID activity. In this research, we have adopted MSTID activity detection criteria based on d-TEC maps, provided that the following conditions are met: That the amplitude of the disturbance exceeds 0.2 TECU (more than 10 times the measurement precision); that the disturbance has at least two clearly defined wave fronts (which guarantees wave behavior); and that the wave pattern persists for at least two consecutive d-TEC maps (which confirms the temporal coherence of the disturbance). The dominant direction of propagation of the detected MSTIDs was also estimated by visually checking the sequential evolution of the wave fronts, classifying it into four sectors: North, East, South, and West for a wave front progression between 315° and 45°, between 45° and 135°, between 135° and 225°, and between 225° and 315°, respectively. The inferred propagation directions were subsequently validated using ionosonde data from EB040, identifying the evolution of the direction of the oblique echoes in the ionograms related to TEC disturbances, as in Paul & Haralambous (2025). We have analyzed the d-TEC maps over the location of EB040 from 2012 to 2016 to identify MSTIDs when SF was observed in the ionograms of EB040. Note that d-TEC maps over Europe are available until September 15, 2017, but we limited our analysis up to December 2016, ensuring that for each year, a full year data-set.

ROTI is an index developed to identify and statistically present small-scale ionospheric irregularities. Pi et al. (1997) defined ROTI as the standard deviation of the rate of TEC. i.e., ROT = (TEC(t) − TEC(t − δt))/δt and $\mathrm{ROTI}=\sqrt{\langle {\mathrm{ROT}}^2\rangle -{\langle \mathrm{ROT}\rangle }^2}$. Based on the “frozen-in” assumption, ionospheric irregularity characteristics are considered stable over short periods. Therefore, fluctuations observed in the ROT can stem from the decrease in the spatial gradient of TEC, denoted as ΔTEC/ΔL, where L represents the horizontal irregularity scale length. In the low and mid-latitude ionosphere, these fluctuations arise from irregularities with smaller scale sizes, ranging from several hundred meters to 2.5 km, contingent upon plasma drift velocities. In short, ROTI maps provided by the DRAWING-TEC project (Tsugawa et al., 2018) use measurements of more than 800 GNSS receivers distributed across Europe that provide data with a temporal resolution δt = 30 s (Otsuka et al., 2013). sTEC time series for each satellite-receiver pair is converted to vTEC whose corresponding IPP is referenced at an altitude of 300 km. The data with satellite elevation angles below 35° are excluded in this procedure. The vTEC are used to compute ROT and ROTI as the standard deviation of time differential TEC in five min (Tsugawa et al., 2018) according to the above formulation. In this way, ROTI values are mapped with a pixel size of 0.15°×0.15° in latitude and longitude and with a temporal resolution of 10 min. Thus, ROTI indicates small-scale ionospheric irregularities and was examined as an indicator of SF activity. In agreement with previous findings, only ROTI values greater than 0.15 TECU/min were considered significant enough to be compared with the SF occurrence (Paul et al., 2024). We have analyzed the ROTI maps over the location of EB040 from 2012 to 2016 to identify significant ROTI activity when SF was observed in the ionograms of EB040. As it occurs with the d-TEC maps, ROTI maps over Europe are available until September 15, 2017. Therefore, we limited our analysis up to December 2016, ensuring that for each year a full year data-set, and we focused on the 5° × 5° geographic grid centered on Roquetes to explore ROTI characteristics during RSF events.

3 Results and discussion

This section presents the results of the analysis of SF occurrence, including diurnal and seasonal patterns, the MSTID-related SF activity and its dependence on solar activity, as well as the potential link between significant ROTI activity and SF occurrence.

3.1 Solar activity dependence on the SF occurrence

We investigated the SF occurrence at EB040 using FSF occurrence results derived from the 1-h scaled foF2. As will be shown in the next section, SF is a phenomenon that occurs predominantly at night at EB040. Thus, for a better characterization of SF climatology, the analysis was restricted to the 16:00–08:00 UT interval. It should be noted that LT and UT closely coincide in the longitude sector of EB040, ensuring that the observed patterns accurately reflect local temporal dynamics. Figure 1a illustrates the evolution of FSF occurrence from 1955 to 2019 and compares it with solar activity over a time interval encompassing six Solar Cycles (19th–24th). The FSF occurrence in Figure 1a represents the ratio of the number of hours in a given year during which FSF is observed in ionograms to the total number of hours in that year for which ionograms have been recorded, both numbers of hours within the 16:00–08:00 UT time window. Figure 1a clearly observes that FSF occurrence maximizes around years of minimum solar activity (e.g., 1957, 1968, 1979, 1989, 2000, and 2014), while years of maximum solar activity (1965, 1975, 1986, 1995, and 2009) show minimum FSF occurrence. These results indicate an inverse dependence of FSF occurrence on solar activity. Moreover, we also evaluated FSF and RSF occurrence rates from 2012 to 2022 (Fig. 1b). Figure 1b shows that the lowest SF activity corresponds to 2014, the maximum of Solar Cycle 24, while the highest SF activity occurs in 2019, the minimum between the end of Solar Cycle 24 and the beginning of Solar Cycle 25. Figure 1b clearly indicates that periods of increasing solar activity (2012–2014 and 2019–2022) show a trend of decreasing SF activity. In contrast, those years with decreasing solar activity (2014–2019) show an increasing trend in the SF occurrence. SF occurrence decreased by 6% from 2012 to 2014, then increased sharply by 23% until 2019, followed by an 18% decline through 2022. This pattern suggests an inverse relationship between SF occurrence and solar activity, consistent with findings from mid-latitude studies (Paul et al., 2018, 2023). Thus, the results depicted in Figure 1 clearly indicate that the SF occurrence is modulated by solar activity with an inverse dependence.

Figure 1 Long-term variation of the SF occurrence at EB040. Panel (a) shows the FSF occurrence rate derived from the 1-h scaled foF2 from 1955 to 2019. Panel (b) shows the occurrence rate of SF, considering both FSF and RSF, from the 5-min ionogram measurements from 2012 to 2022. The dotted red rectangle in Panel (a) represents the same interval (2012–2019) as shown in Panel (b).

An inverse dependence between SF occurrence and solar activity was already established in the literature (Singleton, 1968; Bowman, 1998; Bhaneja et al., 2009; Paul et al., 2019, 2022). Paul et al. (2022) analyzed SF observations over Nicosia (35.0°N, 33.2°E), Cyprus, a low mid-latitude station in the eastern European longitude sector, for the entire Solar Cycle 24, suggesting an inverse solar activity dependence of SF. However, Singleton (1968) reported a higher SF occurrence at higher mid-latitude regions (around 50°N) during periods of high solar activity. It is well established that solar activity modulates the global mean neutral density of the upper atmosphere (e.g., Emmert, 2009) and that MSTID amplitudes at lower mid-latitudes are significantly damped due to higher UA-NPD (Bowman, 1960). Thus, the results of our analysis, based on a very long dataset, support that MSTID’s driven SF events are influenced by solar activity. In fact, the role of UA-NPD in the SF activity may lead to approximately three times higher SF occurrence during years of low solar activity compared to years of high solar activity (Bowman, 1960) which is quite consistent with the peak-to-valley differences in the SF occurrence shown in Figure 1.

3.2 Diurnal spread F occurrence characteristics

We analyzed the day-to-day variation in the occurrence of both FSF and RSF observed in the ionograms from EB040 for the years 2012–2022. Figure 2 depicts the occurrence of FSF and RSF during this period. Each blue solid circle represents an SF event (RSF and/or FSF) observed for at least 15 consecutive minutes, corresponding to three or more consecutive ionograms with a 5-min resolution. The occurrence of SF is practically confined between dusk and dawn, with rare daytime events occurring during the evening hours of the summer months. SF during summer typically manifests from 19:30 to 05:30 UT in EB040, while in winter it shifts slightly later, between 20:30 and 06:30 UT.

Figure 2 Day-to-day occurrence of both FSF and RSF at EB040 as a function of the day of year (DOY) and hour of the day (UT). Note that UT and LT are practically the same at EB040.

Figure 2 shows an evident seasonal variation of SF activity. SF occurrence maximizes at solstices, with a primary peak in June–July, a secondary peak in December–January, and a clear minimum during the equinoxes, in agreement with previous studies in Europe (Paul et al., 2023). This pattern is more noticeable in years of higher solar activity, i.e., 2012–2014, 2021 and 2022 (panels (a)–(c), (j), and (k) of Fig. 2, respectively). The lowest SF occurrence is reported for 2014, which corresponds to the maximum of Solar Cycle 24, while the highest is reported for 2019, the minimum between Solar Cycles 24 and 25.

The results of the analysis of this eleven-year dataset (2012–2022) evidence an inverse dependence of SF occurrence on solar activity, as well as the fact that SF is a nighttime phenomenon at western European mid-latitudes. Moreover, a seasonal pattern is evident, with the largest SF occurrence in June–July, a secondary peak in December–January, and a minimum during the equinoctial months. Other studies have reported similar results at low- to mid-latitude European stations (<50°N) for 2020–2021. Paul et al. (2023) show that SF occurrence peaks in summer, with some activity in winter and a distinct minimum at equinoxes. Paul et al. (2023) report also consistent findings with SF occurrence concentrated between 20:00 and 04:00 LT in summer and 18:00 to 05:00 LT in winter. However, the present work covers a much longer analysis period (2012–2020), which practically spans the entire Solar Cycle 24, making these findings more robust.

3.3 Seasonal pattern of the SF occurrence

We have analyzed the seasonal variation of SF occurrence observed at EB040 for the 11 years 2012–2022, distinguishing between FSF and RSF occurrences. The histograms plotted in Figure 3 depict the seasonal variation of SF, FSF and RSF occurrence rates. Using the same approach applied in Figures 1a, 1b, we have separately identified the hours characterized by SF, RSF and FSF for each month from 2012 to 2022 and calculated the respective occurrence rates by computing the ratio of SF-, RSF- and FSF-affected hours to the total number of hours with ionogram measurements. As in the previous sub-section, we consider SF occurrence (RSF and/or FSF) if it is observed in ionograms for at least 15 consecutive minutes or more (i.e., 3 or more consecutive ionograms with a 5-min sampling interval of soundings).

Figure 3 Seasonal variation of SF occurrence observed from 2012 to 2022. The histograms depict the monthly occurrence rates (1–12: January–December) for both FSF and RSF. The blue dotted line (right y-axis) indicates the solar activity.

The annual distribution of the occurrence rates of SF, FSF and RSF illustrated in Figure 3 shows a clear seasonal pattern for all years from 2012 to 2022. This pattern, which was suggested in Figure 2 for SF, is now clearly evident: SF, FSF, and RSF all have maximum occurrence rates at solstices, with the June–July maximum more pronounced than that of December–January, and a minimum occurrence at equinoxes. SF occurrence in June–July is observed at about 65–70% of the time on average. December–January observes an average SF occurrence for 35–40% of the time. Equinoctial months observe an average SF occurrence for less than 20% of the time. The results in Figure 3 suggest that the SF occurrence rate tends to decrease during periods of increasing solar activity (2012–2014 and 2019–2022), while showing an increasing trend during the declining phase of Solar Cycle 24 (2014–2019). These results agree with those reported in Section 3.1, again indicating that solar activity modulates the occurrence of SF, RSF and FSF, with an inverse dependence. It is also noticeable from Figure 3 that the RSF occurrence rate is higher than the FSF occurrence rate for almost every month under analysis. Specifically, an average of 69% of the observed SF events correspond to the RSF type, while only an average of 33% correspond to the FSF type. This clearly indicates that the type of SF observed at EB040 is dominantly RSF.

This pattern is consistent with that obtained in other studies for mid-latitude stations (e.g., Bowman, 1960; Hajkowicz, 2007; Chen et al., 2011; Paul et al., 2019). The annual variation of SF occurrence shows the largest values for June–July months compared to other season months (e.g., Chen et al., 2011; Paul et al., 2019), with a secondary enhancement of SF occurrence in December–January (Bowman, 1960; Hajkowicz, 2007; Paul et al., 2022). Moreover, a dominant occurrence of RSF compared to FSF has been observed over low- to mid-latitudes during summer months (Paul et al., 2023). A similar pattern has been observed at latitudes below 44°–48°S (Hajkowicz, 2007) in the Southern Hemisphere, albeit based on a limited one-year time interval.

Furthermore, Rastogi & Woodman (1978), through a study combining vertical incidence ionograms with an incoherent scatter radar at equatorial latitudes, observed that echoes due to RSF were significantly stronger than those associated with FSF. They suggested that RSF is caused by very strong and irregular vertical gradients of ionization at the base of the F-region, including structures down to the 3 m scale length. However, equatorial irregularities causing SF occurrence are dominantly linked to equatorial plasma bubbles (EPB) (e.g., Booker & Wells, 1938; McNamara et al., 2013), which are an unusual phenomenon at mid-latitudes (e.g., Cherniak & Zakharenkova, 2016; Campuzano et al., 2023). In general, the consistent dominance of RSF over FSF observed at EB040, as well as their very similar seasonal variation, which is also reported in other regions and studies, underlines the significance of RSF over low mid-latitudes during the summer months. These findings highlight the influence of seasonal and solar activity variability on SF. Mid-latitude SF is often driven by wave-like density perturbations originating from the bottomside F-layer and interacting with horizontal plasma gradients (Bowman, 1990). F-layer uplift, sporadic-E layers, and E–F region electrodynamic coupling – often triggered by TIDs – have also been proposed as key drivers of SF (Tsunoda & Cosgrove, 2001). The seasonal SF pattern aligns with the seasonal MSTID pattern over Europe, especially during summer (Kotake et al., 2006; Otsuka et al., 2013; Paul et al., 2023). MSTIDs enhance plasma instabilities through processes like Perkins instability, supported by eastward neutral winds and background electric field in the post-sunset period. Thus, the seasonal and interannual patterns observed at EB040 suggest that nighttime mid-latitude SF is modulated by both solar activity and MSTID-related wave-plasma interactions, emphasizing the role of electrodynamic processes in shaping ionospheric variability.

3.4 SF-MSTID link

This section shows the analysis of SF occurrence linked with the MSTID activity at around the EB040 ionosonde location. For such a purpose, we use MSTID activity indicators derived from the GNSS d-TEC maps from 2012 to 2016, focusing on the time intervals when SF is observed at EB040. Saito et al. (1998a) were the first researchers to reveal MSTID activity from d-TEC maps in Japan using data from a dense GNSS receiver network. Subsequent studies have confirmed that MSTID activity can be inferred from d-TEC maps (Otsuka et al., 2013, 2021). To illustrate an example, Figure 4 displays several SF events recorded over Roquetes along with simultaneous MSTIDs signatures evident on the d-TEC maps. Following the approach of Otsuka et al. (2013), we have identified the occurrence and propagation direction of MSTIDs events associated with SF detected by the EB040 ionosonde.

Figure 4 Examples of d-TEC maps (left plots) indicating signatures of MSTIDs jointly with ionograms recorded at Roquetes (right plots), simultaneous with MSTID activity at indicated timestamps. The pink location marker in the d-TEC maps indicates the position of the EB040 station. See text for details.

The d-TEC maps in Figures 4a–4e show MSTID signatures over the ionosonde location (indicated by a black circumference) and the propagation direction of the disturbance (represented by dotted pink arrows) along with the corresponding ionograms recorded on January 27, 2012; June 4, 2013; May 24, 2014; January 13, 2014; and January 5, 2016 respectively. Although not shown here, a detailed analysis of EB040 ionograms confirms SF activity from January 27, 2012 at 19:45 UT to January 28 at 07:00 UT. Simultaneously, the d-TEC maps reveal MSTID activity from 19:00 UT on January 27 to 02:00 UT on January 28. We included a supplementary video (Vid1) to more clearly demonstrate the temporal and spatial correspondence between MSTID activity and the observed SF signatures during this event. This video captures a representative interval from 21:50 UT to 22:40 UT on January 27, displaying the sequential evolution of d-TEC perturbations together with the associated ionograms. The mapped perturbations exhibit a coherent southward propagation pattern across the observation region, which is temporally aligned with the occurrence of RSF structures in the ionograms. This propagation direction is further supported by the presence of oblique echoes – highlighted using colour-coded annotations in the ionograms – serving as additional observational evidence for the directionality of the underlying ionospheric disturbances (Paul & Haralambous, 2025), consistent with the morphology observed in Figure 4a and Vid1. A similar analysis was carried out for each of the additional events shown in Figures 4b through 4e. In all cases, the d-TEC maps exhibit clear MSTID activity with dominant southward propagation, coinciding in both time and location with SF occurrences recorded at EB040. To support and visualize these results, we provide supplementary video material (Vid2–Vid5), corresponding to each event. These videos illustrate the propagation dynamics of MSTID structures and their association with the onset and development of SF signatures in the ionograms.

We evaluated the propagation direction of SF-associated MSTID events over Roquetes for the period 2012–2016 as described in Section 2.2. The results are summarized in the statistical histogram shown in Figure 5, which evidences that a vast majority (approximately 94%) of the SF-associated MSTIDs propagated southward. This predominant southward propagation agrees with earlier findings (e.g., Miller et al., 1997; Kelley & Miller, 1997), which suggest that such directional preference cannot be fully explained by classical AGW theory alone. Instead, electrodynamical mechanisms – such as Perkins instability and E–F region coupling – are thought to play a significant role.

Figure 5 Bar plot of propagation directions for the SF-associated MSTIDs over Roquetes from 2012 to 2016. The x-axis represents the observation year, and the y-axis stands for the occurrence percentage (%) of propagation directions.

We have comprehensively investigated the MSTIDs activity observed in the d-TEC maps over the location of the EB040 ionosonde in conjunction with SF occurrence for the time interval spanning the years 2012–2016. Thus, we performed a statistical analysis to determine the occurrence of SF together with the existence of MSTIDs. For this purpose, we considered only those MSTID events whose d-TEC values exhibit fluctuations larger than 0.2 TECU (Otsuka et al., 2013). The histograms depicted in Figure 6 show the annual distribution of the ratio of hourly SF occurrence for each month, along with the ratio of simultaneous hourly MSTID activity associated with the SF for the years 2012–2016. The results of Figure 6 clearly show that MSTID activity linked to SF occurrence follows the same annual trend observed for the SF occurrence, with maximum occurrence during the solstices (more pronounced in June–July than in December–January), and minimum occurrence during the equinoxes. MSTID activity linked to the SF occurrence is quite similar for the years 2012 and 2013 (Figs. 6a and 6b respectively), which also share very similar solar activity conditions. However, MSTID activity associated with SF occurrence tends to decrease with increasing solar activity (from 2013 to 2014), and it increases as solar activity decreases (from 2014 to 2016). In other words, MSTID activity linked to SF is influenced by solar activity, exhibiting an inverse dependence. These results underscore the potential correlation between MSTIDs and SF occurrences and highlight the influence of solar activity on both MSTID activity and SF occurrence.

Figure 6 Seasonal variation of SF occurrence and associated MSTID activity for the years 2012–2022. Histograms depict the monthly occurrence rates (1–12: January–December) for both SF and MSTIDs. The blue dotted line shows solar activity (right y-axis).

We also analyzed the onset time of SF events observed at EB040 from 2012 to 2016 and compared them with the onset times of MSTID events (identified in the d-TEC maps) over the ionosonde location. Figure 7 illustrates the scatter plots of MSTID events onset times as a function of SF events onset times for the years 2012–2016. The blue line in plots 7a–7e represents the best linear fit of each scatter plot, and the corresponding linear correlation coefficient is shown in the upper left corner. Figure 7 evidences a high correlation, with a determination coefficient exceeding 0.96, between the onset of SF occurrence and that of MSTID activity. This finding suggests that most of the SF events at EB040 are mainly driven by MSTID activity.

Figure 7 Temporal relationship between MSTID activity onset and SF occurrence onset for the years 2012–2016.

3.5 SF occurrence during enhanced ROTI activity

The SF occurrence results described in the previous analysis show that both SF types, RSF and FSF, are commonly associated with MSTIDs. Results also evidence that station EB040 is more prone to RSF than FSF. In addition, RSF is often accompanied by small-scale ionospheric irregularities (Bowman, 1990), which can be effectively captured through enhanced values of the ROTI. As already mentioned above, Paul et al. (2024) demonstrated a strong link between significant ROTI activity and RSF occurrence at European mid-latitudes. This section shows the analysis of RSF occurrence linked with enhanced ROTI activity (ROTI > 0.15 TECU/min) over EB040 to better understand the impact of the small-scale size ionospheric irregularities on the RSF and of potential RSF drivers.

For this purpose, we used ROTI maps over Europe provided by the DRAWING-TEC project (Tsugawa et al., 2018) from 2012 to 2016, and investigated RSF activity over EB040. Figures 8 and 9 show several examples of RSF events observed at EB040 coincident with enhanced ROTI activity under disturbed and quiet geomagnetic conditions, respectively. Figure 8 depicts ROTI maps with a pink location marker inside a pink circumference indicating the location of EB040, jointly with the corresponding ionograms recorded on June 29, 2013; June 22, 2015; and August 23, 2016. June 29, 2013 was a geomagnetically disturbed day (Dst ≤ −100 nT), and EB040 observed RSF coinciding with ROTI > 0.25 TECU/min from 00:20 to 02:00 UT. Panels (b) and (c) of Figure 8 show similar intense ROTI events associated with RSF that occurred during a strong geomagnetic storm (reaching Dst ≤ −110 nT) and a moderate geomagnetic storm (reaching Dst ≤ −70 nT) respectively.

Figure 8 Examples of ROTI maps (left plots) with significant ROTI activity at around the ionosonde location (pink circumference), jointly with ionograms recorded at Roquetes (right plots), simultaneous with significant ROTI activity at indicated timestamps. The pink location marker in ROTI maps indicates the position of the EB040 station. See text for details.

Figure 9 As in Figure 8 but for indicated quiet-time intervals.

In parallel with Figures 8 and 9 depicts ROTI maps and corresponding ionograms showing RSF events that occurred during significant ROTI activity under geomagnetically quiet conditions. On May 25, 2012 (Dst ≥ −17 nT), EB040 observed RSF along with ROTI > 0.2TECU/min from 01:00 to 03:00 UT. Panels (b)–(e) of Figure 9 display ionograms showing similar RSF activity in conjunction with significant ROTI activity under quiet geomagnetic conditions (Dst > 25 nT). The key distinction between the geomagnetically active and quiet periods, as illustrated in Figures 8 and 9, lies in the magnitude of the ROTI fluctuations. ROTI values were noticeably higher during geomagnetically active conditions, ranging from approximately 0.25–0.3 TECU/min. In contrast, ROTI values over Roquetes were comparatively lower during quiet periods, typically ranging from 0.15 to 0.2 TECU/min.

We have comprehensively investigated the significant ROTI activity observed over the location of the EB040 ionosonde from the ROTI maps in conjunction with RSF occurrence for the period 2012–2016. A statistical analysis was performed to determine the co-occurrence of RSF and significant ROTI activity, taking into account only those events with ROTI > 0.15 TECU/min.

The histograms depicted in Figure 10 show the annual distribution of the ratio of hourly RSF occurrence in the given month, along with the annual distribution of the ratio of hourly simultaneous occurrence of significant ROTI activity associated with RSF from 2012 to 2016. Figure 10 indicates that significant ROTI activity potentially linked to RSF occurrence exhibits an evident annual variation, with maximum occurrence at the summer solstice. However, the secondary maximum during the winter solstice months, which is evident in the RSF occurrence, is not clearly present in the significant ROTI activity linked to RSF. Moreover, Figure 10 evidences that there is a higher percentage of simultaneous ROTI activity and RSF occurrence during years of higher solar activity. In addition, comparing the results of Figures 6 and 10, there is a lower percentage of simultaneous significant ROTI activity and RSF occurrence compared to simultaneous MSTID activity and SF occurrence. Overall, Figure 10 shows that the distribution of RSF events linked to significant ROTI activity is characterized by a summer peak and by a direct influence of the solar activity, with a larger occurrence of events during periods of higher solar activity.

Figure 10 Annual variation of RSF occurrence and of significant ROTI activity linked to RSF observed for the years 2012–2022. Histograms depict the monthly occurrence rates (1–12: January–December) for both RSF and ROTI activity. The blue dotted line indicates solar activity (right y-axis).

Figure 11 illustrates the solar activity impact on RSF events linked to significant ROTI activity. The histogram in Figure 11a depicts the ratio of the RSF events linked to significant ROTI activity with respect to the total RSF events observed for the years under analysis, and it demonstrates an evident solar activity dependence, with a larger ratio of RSF linked to significant ROTI during years of higher solar activity. To further investigate the possible conjunction of RSF activity with significant ROTI activity, we classified RSF events into four categories according to their duration. The histograms of Figure 11b depict the ratio of the RSF events linked to significant ROTI activity with the indicated duration with respect to the total RSF events of that duration observed for the years under analysis. Category 1 accounts for RSF events linked to significant ROTI activity with a duration of less than 30 min, category 2 accounts for RSF events with a duration from 30 min to 1 h, category 3 accounts for RSF events with a duration from 1 to 2 h, and category 4 accounts for RSF events with a duration longer than 2 h. According to Figure 11b, most of the RSF events observed over EB040 with a duration longer than 2 h (83%) are related to significant ROTI activity. The percentage of RSF linked to significant ROTI activity decreases as RSF events have shorter durations. Therefore, this indicates that significant ROTI activity events are related to long-standing RSF events at the mid-latitude station of EB040. Figure 11c depicts the scatter plot of the ratio of the RSF events linked to significant ROTI activity with respect to the total RSF events observed for the years under analysis as a function of solar activity. The solid line in Figure 11c represents the best linear fit of the scatter plot, and the upper left corner shows the corresponding linear determination coefficient. Figure 11c indicates that only 62% of the RSF observed by the EB040 ionosonde from 2012 to 2016 are related to intense ROTI events. Moreover, the RSF events, which are synchronized with significant ROTI events, show a positive correlation (with a high correlation coefficient >0.97) with solar activity.

Figure 11 Solar activity impact on RSF events linked to significant ROTI activity. See text for details.

4 Summary and conclusion

This research has addressed in detail the climatology of the nighttime SF characteristics at the mid-latitude ionospheric station EB040 in Roquetes, Spain, aiming at providing a physical explanation for the mechanisms linking SF occurrence and MSTID activity. We have conducted a comprehensive long-term study, encompassing the years 1955–2019, based on SF identification in ionogram scaled by experts from the EB040 ionosonde to demonstrate the influence of solar activity on SF at mid-latitude. The obtained results show that solar activity modulates the occurrence of FSF with an inverse dependence, i.e., FSF occurrence increases as solar activity decreases, and vice versa. Furthermore, we have performed a detailed analysis to investigate yearly, seasonal, and diurnal variations of SF for both FSF and RSF using the ionograms recorded from 2012 to 2022 at EB040. Results of this analysis evidence that SF is a night-time phenomenon at western European mid-latitudes and indicate a similar solar activity dependence. The annual distribution of the occurrence rate of both SF types, FSF and RSF, shows a clear seasonal pattern, with a maximum occurrence rate at solstices and a minimum occurrence at equinoxes. The maximum SF occurrence rate observed in June–July is larger than that in December–January. Moreover, there is clear evidence that the rate of occurrence of RSF is higher than that of FSF, since 69% of the observed SF events correspond to the RSF type, while only 33% of the observed SF events correspond to the FSF type. We have to note that it is possible to simultaneously identify RSF and FSF features in the same ionogram, which is known as MSF (mixed RSF and FSF). Although these MSF events provide valuable insight into complex ionospheric dynamics, they are relatively infrequent in our dataset, and specific analysis on MSF events was not conducted in this study. For the purposes of our statistical analysis of SF occurrences, each MSF event has been counted as one RSF and one FSF event, allowing us to include their contribution while maintaining consistency in event classification.

The seasonal dependence of SF events over EB040, peaking during the summer months, aligns well with earlier studies. Paul et al. (2024) reported a similar trend across European low mid-latitude stations (latitude < 45°N), where SF events were found to peak between May and July, with an average occurrence rate of approximately 67%. Paul et al. (2019, 2022) also observed a primary SF occurrence peak from May to August (~32.5%) and a secondary peak during January–February (~20%) over the eastern Europe mid-latitude station in Nicosia for the period 2009–2020. Supporting this, Hernández-Pajares et al. (2006) reported enhanced nighttime MSTID activity over Europe near the June solstice, based on GNSS TEC analysis – corresponding to increased SF occurrence in low mid-latitudes. Similarly, Huang et al. (2011) identified higher SF activity during summer and winter at two mid-latitude stations in China. Chen et al. (2011) also noted that MSTIDs frequently occur during both solstices, particularly in summer. Bowman (1992) linked this seasonal pattern to variations in UA-NPD, which exhibit maxima during equinoxes and minima during solstices. Since both GW amplitude and the linear growth rate of the Perkins instability are inversely proportional to neutral density (Hines, 1963; Perkins, 1973), summer conditions in the Northern Hemisphere may favour the development and amplification of MSTIDs. This, in turn, increases the likelihood of SF formation. Furthermore, MSTID characteristics may differ between solstices due to variations in perturbation strength and the ease of irregularity generation, especially during summer. While MSTIDs promote ionospheric instability, the formation of irregularities also depends on additional factors, including the background electric field, neutral winds, and the vertical structure of the ionosphere (Saito et al., 2001).

In addition, further analysis of different indicators of ionospheric irregularities, including indicators of MSTIDs (d-TEC and ROTI maps), for the years 2012–2016 has been investigated to assess the MSTID-driven SF occurrence and its dependence on solar activity, as well as the potential link of the significant ROTI events with RSF occurrence at mid-latitudes. The obtained results evidence that MSTID activity linked to the SF occurrence follows the same annual trend as that of SF occurrence, with occurrence peaks at solstices, June–July being larger than December–January, and minimum occurrence in the equinoxes. There is an indication that the MSTID activity linked to the SF is influenced by solar activity following an inverse dependence, and that about 85% of SF events observed in EB040 were driven by MSTID activity, with a high correlation (exceeding 0.96) between the onset of the SF occurrence and that of the MSTID activity. The results also show that significant ROTI activity is linked to RSF occurrence, which exhibits an evident annual variation, with maximum occurrence in the summer solstice, and which is influenced by solar activity with a positive correlation (>0.97). Moreover, significant ROTI activity events are related to long-standing RSF events in the mid-latitude station of EB040, and only about 60% of the RSF observed by the EB040 ionosonde are related to significant ROTI values.

The high correlation (~0.9) between the onset times of SF and MSTID events over EB040, together with the coincident annual and long-term variations of SF occurrence obtained in this investigation, leads us to conclude that MSTIDs are the main cause of mid-latitude SF in this region. Bowman (1990) noted that SF events are typically preceded by TID wave trains, which induce significant vertical plasma displacement in the F-region. These include an abrupt layer uplift – often preceded by a brief suppression – with periodicities around 20 min. This vertical displacement results in a decrease in the F-layer peak electron density during the uplift, followed by a recovery of the layer to its original height. The associated wavelike tilts in electron density cause off-vertical reflections in ionograms, manifested as SF once the structures drift sufficiently from the zenith (Bowman & Monro, 1988). The sloping gradients induced by the MSTID-driven uplift enhance the tilts of F-region plasma structures, which become visible in ionograms as SF, particularly during the recovery phase of the layer. This behaviour is consistent with early studies (McNicol et al., 1956) and confirms the role of macroscale height fluctuations in amplifying small-scale irregularities that generate SF (Bowman, 1990). Observations of MSTIDs over mid-latitudes show a dominant direction of propagation from north to south at speeds of 40–100 m/s (Bowman, 1981; Otsuka et al., 2011, 2013), often without significant attenuation, supporting SF generation. More recent studies reported that nearly 92.5% of SF events during the European summer (2020–2022) were linked to MSTIDs (Paul et al., 2023). Solar activity also plays a crucial role in modulating MSTID occurrence and thus SF. MSTID activity is inversely related to solar activity, with more frequent MSTIDs during solar minima (Candido et al., 2008; Yu et al., 2016; Otsuka et al., 2021). This is likely due to reduced background ionospheric conductivity during solar minima, which favours the development of the Perkins instability – a key mechanism in MSTID growth (Kotake et al., 2007). The linear growth rate of this instability is inversely proportional to the UA-NPD, which varies semiannually with minima at solstices (Hines, 1963), and it is modulated by the solar activity cycle (e.g., Emmert, 2009). Consequently, MSTID amplitudes at lower mid-latitudes are significantly damped due to higher UA-NPD. This seasonal and Solar Cycle modulation gives rise to conditions that are highly favourable to MSTID-driven SF for solstice nights and periods of low solar activity.

The observed southward propagation of nighttime MSTIDs over Europe, including the latitude range of EB040, offers important indications about their generation mechanisms. According to Otsuka et al. (2013), the preferred alignment of MSTID wavefronts along the NW–SE direction and their dominant southwestward phase velocity are difficult to reconcile with classical AGW theory alone. Instead, these characteristics align more consistently with electrodynamically driven instabilities. The Perkins instability (Perkins, 1973) has long been considered a candidate mechanism due to its preference for similar wavefront alignments; however, its intrinsic linear growth rate is too small to account for the amplitude and prevalence of MSTIDs observed at mid-latitudes (Kelley et al., 2002). This limitation has led to the hypothesis that coupling between the E- and F-regions – particularly involving sporadic E (Es) layers – can enhance instability growth through polarization electric fields, creating a positive feedback loop that facilitates MSTID development (Tsunoda & Cosgrove, 2001; Kelley et al., 2003). Observations by Otsuka et al. (2013) of wavelike Es structures co-aligned with MSTID wavefronts support this mechanism in the European sector. Beyond local generation, Kelley (2011) proposed that high-latitude sources may also play a role, particularly through the propagation of large-scale AGWs into midlatitudes. Their simulations and observations suggest that AGWs generated in the auroral zone, possibly via structured Joule heating, can propagate equatorward if they align along directions of minimal Joule damping – specifically those predicted by the Perkins instability for unstable growth. In this context, damping is minimized when the perturbation electric field and neutral wind-driven current are nearly cancelled by the polarization electric field-induced current, reducing the overall Lorentz forcing – caused by the interaction between the electric current and the geomagnetic field – and consequently minimizing the Joule heating that would otherwise damp the wave (Kelley, 2011). Under such conditions, the ion drift (E × B/B²) becomes nearly parallel to the neutral wind in the wave, resulting in weak dissipation and allowing the waves to propagate to lower latitudes relatively unattenuated. Supporting the weak-damping hypothesis, DE-2 satellite measurements demonstrated that the propagation direction of MSTIDs is often aligned with the E × B drift of the plasma (Saito et al., 1998b), indicating minimal differential motion between the wave and the background ionosphere. Importantly, only AGWs with sufficiently long vertical wavelengths can survive viscous damping in the lower thermosphere and reach the F-region, as short-wavelength components are strongly attenuated (Kelley, 2011). Once present in the F-region, these waves may act as seed perturbations, which are then amplified through electrodynamic coupling processes. Thus, the MSTIDs responsible for midlatitude SF over EB040 likely originate from a hybrid mechanism: Local generation via E–F region electrodynamical coupling and the Perkins instability, as well as selective propagation of high-latitude gravity waves that exploit favourable damping conditions. The dominance of southward propagation and the spatial-temporal coincidence with Es activity suggest that in situ coupling processes are primary drivers, but external forcing from high-latitude sources may act as important triggers or amplifiers under appropriate background conditions.

The correlation between SF and ROTI events provides further evidence of electrodynamic drivers of SF activity. Paul et al. (2024) found that all RSF occurrences in 2011 were associated with ROTI values exceeding 0.15 TECU/min. Although not all highly significant ROTI events result in SF, the temporal alignment supports their shared common physical basis. ROTI activity peaks during summer over lower mid-latitudes and increases with solar activity due to enhanced ionospheric density and gradient-driven instabilities (Liu & Wu, 2021). Our results show that significant ROTI activity linked to RSF occurrence exhibits an evident annual variation, with maximum occurrence in the summer solstice, and it is influenced by the solar activity with a positive correlation (>0.97). Moreover, significant ROTI events are related to long-standing RSF events at the mid-latitude station of EB040. However, the absence of a clear secondary maximum in the significant ROTI activity linked to RSF for winter solstice months, in contrast to the secondary maximum observed for SF occurrence in conjunction with MSTID activity, together with the direct influence of solar activity, with a larger occurrence of events during periods of higher solar activity, indicate additional RSF drivers distinct from MSTID activity, especially for those long-lasting RSF events. We can speculate that strong ionospheric irregularities related to enhanced solar activity events, such as storm-driven ionospheric irregularities (e.g., Blanch et al., 2005; Campuzano et al., 2023; Calabia et al., 2024), may play a role.

From this extensive long-term investigation over EB040, we can conclude that MSTIDs generated by the AGW through the Perkins instability are the most likely cause of mid-latitude SF occurrence, based on the following main results:

1. The occurrence of SF exhibits an inverse dependence on solar activity. This inverse relationship can be explained by the direct correlation between UA-NPD and solar activity. UA-NPD increases as solar activity increases and strengthens the damping of MSTIDs.

2. SF is a night-time phenomenon at western European mid-latitudes, typically occurring from 20:00 to 05:00 UT, and the diurnal SF occurrence pattern is influenced by solar activity.

3. SF occurrence exhibits a clear semiannual distribution, peaking at solstices with higher occurrence in June–July than in December–January, and minimizing at equinoxes.

4. SF occurrence is highly correlated with MSTID activity which shows the same clear semiannual distribution as SF occurrence.

5. Significant ROTI activity linked to the RSF occurrence is influenced by the solar activity with a positive correlation; it exhibits an annual variation, with maximum occurrence at the summer solstice, and significant ROTI events are related to long-lasting RSF.

Acknowledgments

This research was carried out as part of the TNA activity hosted in remote access at the Ebro Observatory Node (Observatorio del Ebro Fundación) under the PITHIA-NRF¹¹ project. The PITHIA-NRF project has received funding from the European Union’s Horizon 2020 research and innovation programme under Grant Agreement No. 101007599. We extend our gratitude to the DRAWING-TEC¹² project and to the Royal Observatory of Belgium for making available the d-TEC and ROTI maps, and the average sunspot numbers respectively. The editor thanks Giorgio Picanço and an anonymous reviewer for their assistance in evaluating this paper.

Funding

This research has been supported by EU Projects PITHIA-NRF (GA 101007599) and T-FORS https://t-fors.eu/ (GA 101081835), and by funds of the Departament de Recerca de la Generalitat de Catalunya and of the Universitat Ramon Llull. Additional support was provided by the project “Towards A Global Model for Spread F” – TAGMOS-F (Ref No: EXCELLENCE/0524/0526), co-funded by the Republic of Cyprus and the European Regional Development Fund (ERDF) through the “EXCELLENCE HUBS” Programme, under the Operational Program “ΘΑλΕΙΑ” 2021–2027.

Data availability statement

Digital ionograms recorded at station EB040 are available at the DIDBASE https://giro.uml.edu/didbase/. Sunspot number is available at SILSO https://www.sidc.be/SILSO/datafiles. Maps of d-TEC and ROTI are available at DRAWING-TEC project.

Supplementary material

Vid1. Temporal and spatial correspondence between MSTID activity and the observed SF signatures during the event observed on 27 January 2012 depicted in Figure 4a.

Vid2. Temporal and spatial correspondence between MSTID activity and the observed SF signatures during the event observed on 4 June 2013 depicted in Figure 4b.

Vid3. Temporal and spatial correspondence between MSTID activity and the observed SF signatures during the event observed on 24 May 2014 depicted in Figure 4c.

Vid4. Temporal and spatial correspondence between MSTID activity and the observed SF signatures during the event observed on 13 January 2015 depicted in Figure 4d.

Vid5. Temporal and spatial correspondence between MSTID activity and the observed SF signatures during the event observed on 5 January 2016 depicted in Figure 4e.

References

All Figures

	Figure 1 Long-term variation of the SF occurrence at EB040. Panel (a) shows the FSF occurrence rate derived from the 1-h scaled foF2 from 1955 to 2019. Panel (b) shows the occurrence rate of SF, considering both FSF and RSF, from the 5-min ionogram measurements from 2012 to 2022. The dotted red rectangle in Panel (a) represents the same interval (2012–2019) as shown in Panel (b).
In the text

	Figure 2 Day-to-day occurrence of both FSF and RSF at EB040 as a function of the day of year (DOY) and hour of the day (UT). Note that UT and LT are practically the same at EB040.
In the text

	Figure 3 Seasonal variation of SF occurrence observed from 2012 to 2022. The histograms depict the monthly occurrence rates (1–12: January–December) for both FSF and RSF. The blue dotted line (right y-axis) indicates the solar activity.
In the text

	Figure 4 Examples of d-TEC maps (left plots) indicating signatures of MSTIDs jointly with ionograms recorded at Roquetes (right plots), simultaneous with MSTID activity at indicated timestamps. The pink location marker in the d-TEC maps indicates the position of the EB040 station. See text for details.
In the text

	Figure 5 Bar plot of propagation directions for the SF-associated MSTIDs over Roquetes from 2012 to 2016. The x-axis represents the observation year, and the y-axis stands for the occurrence percentage (%) of propagation directions.
In the text

	Figure 6 Seasonal variation of SF occurrence and associated MSTID activity for the years 2012–2022. Histograms depict the monthly occurrence rates (1–12: January–December) for both SF and MSTIDs. The blue dotted line shows solar activity (right y-axis).
In the text

	Figure 7 Temporal relationship between MSTID activity onset and SF occurrence onset for the years 2012–2016.
In the text

	Figure 8 Examples of ROTI maps (left plots) with significant ROTI activity at around the ionosonde location (pink circumference), jointly with ionograms recorded at Roquetes (right plots), simultaneous with significant ROTI activity at indicated timestamps. The pink location marker in ROTI maps indicates the position of the EB040 station. See text for details.
In the text

	Figure 9 As in Figure 8 but for indicated quiet-time intervals.
In the text

	Figure 10 Annual variation of RSF occurrence and of significant ROTI activity linked to RSF observed for the years 2012–2022. Histograms depict the monthly occurrence rates (1–12: January–December) for both RSF and ROTI activity. The blue dotted line indicates solar activity (right y-axis).
In the text

	Figure 11 Solar activity impact on RSF events linked to significant ROTI activity. See text for details.
In the text