Unseasonal development of post-sunset F-region irregularities over Southeast Asia on 28 July 2014: 2. Forcing from below?

The development of Equatorial Spread-F (ESF)/Equatorial Plasma Bubble (EPB) irregularities in the nighttime equatorial ionosphere has been studied for several decades since it was first reported by Booker and Wells (1938). From the detailed studies by several researchers, the general morphological features of EPBs such as the occurrence variability with local time, season, latitude, longitude, solar cycle, and geomagnetic activity were more or less thoroughly understood (Abdu et al. 1981; Kelley and McClure 1981; Tsunoda 1985; Rama Rao et al. 2005; Tulasi Ram et al. 2008; Kumar et al. 2016). However, the anomalous day-to-day variability in the occurrence of EPBs still remains to puzzle the researchers over the years. Because of their potential impact on trans-ionospheric radio communications and satellite-based navigation systems, EPBs have drawn wide attention and become one of the important topics of space weather interest today (Kelley et al. 2011; Woodman 2009).

The ESF/EPB irregularities generally initiate at the bottom side F-layer through the Rayleigh-Taylor (RT) instability (Haerendel 1973) and grow nonlinearly to the topside ionosphere via polarization electric fields within them. The pre-reversal enhancement (PRE) of the equatorial zonal electric field and steep vertical density gradient due to bottom side recombination sets the conditions most favorable for the development of EPBs during the post-sunset hours (Farley et al. 1986; Tulasi Ram et al. 2006). Further, the spatial large-scale wave structure (LSWS) at the base of F-layer is believed to be the potential seed for the development of EPBs and determines the location of bubble onset (Tsunoda and White 1981; Tsunoda 2005; Tsunoda et al. 2011; Tulasi Ram et al. 2012, 2014). Besides the PRE and initial seeding by LSWS, the symmetrical plasma distribution in the conjugate hemispheres and meridional neutral winds also play significant roles for determining the onset of EPBs (Maruyama and Matuura 1984; Maruyama 1988).

Among the above necessary favorable conditions, the post-sunset height rise (PSSR) of equatorial F-layer via the PRE is known to be the most important factor. It is also often shown that the EPBs would always develop when the equatorial F-layer is elevated to sufficiently higher altitudes (Farley et al. 1970; Rastogi 1980; Kelley and Maruyama 1992; Fejer et al. 1999; Anderson et al. 2004; Tulasi Ram et al. 2006, 2007; Li et al. 2007; Huang and Hairston 2015). However, the strength of the PRE (and PSSR) itself exhibits significant day-to-day variability (Scherliess and Fejer 1999), thereby leaving the early prediction of EPBs far from the reality. The PRE is generally large during the seasons when the sunset terminator is closely aligned with the local magnetic meridian (Abdu et al. 1981; Tsunoda 1985). The close alignment of the sunset terminator with the magnetic meridian causes simultaneous decay of E-region conductivity at both ends of the field line and maximizes the strength of the PRE. Therefore, over Indian and Southeast Asian longitudes where the magnetic declination is close to zero, the PRE strength maximizes during equinoxes and becomes weaker during solstices (e.g., Tulasi Ram et al. 2006). As a result, the occurrence of post-sunset EPBs maximizes during the equinoxes and becomes lowest during the solstices, particularly during the June solstices over Indian and Southeast Asian sector (e.g., Retterer and Gentile 2009; Otsuka et al. 2009; Ajith et al. 2015). There can be another class of EPBs that develop around post-midnight hours of June solstices over Indian and Southeast Asian sector which have different characteristics compared to post-sunset EPBs (Otsuka and Ogawa 2009; Dao et al. 2016; Ajith et al. 2016). However, the present investigation is focused only on the post-sunset EPBs that are associated with PRE and PSSR.

In contrast to the generally known morphology, the PRE may be occasionally suppressed during equinoxes and, conversely, can be enhanced during solstices, thereby instituting enigmatic day-to-day variability even under quiet geomagnetic conditions. Such unseasonal cases of PRE suppression during equinoxes and enhancement during solstices are difficult to predict. Despite some recent success in using coupled thermosphere-ionosphere modeling to describe day-to-day EPB variability during equinox periods (e.g., Carter et al. 2014a, 2014b, 2014c), EPB occurrence during solstices is not well predicted by either theoretical or climatological models, and even by the more sophisticated self-consistent Thermosphere-Ionosphere Electrodynamics General Circulation Model (TIEGCM) (e.g. Carter et al. 2014c). Investigating such unseasonal cases in greater detail can advance our understanding on the possible factors responsible for the day-to-day variability in the PRE and EPB occurrence. In the present paper (and in a companion paper by Carter et al. 2018), we report and investigate an interesting case of unseasonal enhancement of PRE and development of EPBs during the post-sunset hours on July 28, 2014, over a wide sector of Indian and Southeast Asian longitudes.

From the results presented in Figs. 1, 2, and 3, it is evident that this unseasonal ESF/EPB event on 28 July 2014 is a large-scale plasma bubble activity spanning over a wide longitudinal sector from India, Indonesia, and Southeast Asian region. Carter et al. (2018, the companion paper) have also shown the widespread L-band scintillation activity over the longitudes of Indian and Southeast Asian region from COSMIC/Formosat-3 GPS-Radio Occultation observations. Further, Fig. 4 indicates that the EPB activity is completely absent over the Indonesian sector during the entire month of July 2014 except this singular event day (28th). Also, the VHF scintillation observations from Kolkata (India) and Chung-Li (Taiwan) exhibit no significant post-sunset scintillations other than on July 28 for the entire month (not shown in the figure). Carter et al. (2018) have already shown this unseasonal ESF event on 28 July 2014 is not associated with any disturbed geomagnetic activity. Further, this unseasonal ESF event is preceded by a substantial elevation of equatorial F-layer (h′F in Fig. 5a, b) compared to its monthly mean variation and day-to-day variability represented by the standard deviation (error bars) at both Indian and Indonesian sectors. This elevation of equatorial F-layer is mainly due to the large pre-reversal enhancement (PRE) in the zonal electric field, which is generally unexpected in the month of July due to the misalignment between the sunset terminator and magnetic meridian (Abdu et al. 1981; Tsunoda 1985). Hence, the causes behind the unusual enhancement of PRE on this event day remain to be the most intriguing question to be addressed here. Though the exact mechanisms responsible for the enhanced PRE are not clearly known, we put forth our views about the possible mechanisms hereunder.

There are three mechanisms generally proposed to explain the PRE in the literature: (i) the curl-free mechanism (CFM) by Rishbeth (1973) and Eccles (1998), (ii) Hall current divergence mechanism (HCDM) by Farley et al. (1986), and (iii) the EEJ feeding to PRE by Haerendel and Eccles (1992). An excellent review on the three competing mechanisms and their relative importance can be found in Eccles et al. (2015). The CFM proposed by Rishbeth (1971, 1973) and Eccles (1998) indicates that the PRE magnitude is proportional to the longitudinal gradient in the downward polarization electric field generated by the F-region neutral wind dynamo. The HCDM by Farley et al. (1986) depends on the Hall conductivity gradient in the off-equatorial E-region and the associated divergence of Hall current which in turn is driven by the downward polarization electric fields generated by the F-region dynamo. Nevertheless, both the CFM and HCDM suggest that the downward polarization electric fields generated by the F-region dynamo region and the simultaneous decay of corresponding off-equatorial E-region conductivity hold the key for the PRE. Tsunoda (1985) has further shown that the rate of decay of total field-line-integrated E-region conductivity (combined E-region conductivities of both hemispheres) controls the magnitude of the PRE. The slower decay of off-equator E-region density in the summer hemisphere reduces the PRE magnitude during the solstices by short-circuiting the polarization electric fields generated by the F-region dynamo (Tsunoda 1985). Abdu et al. (2004) have shown that the PRE is very sensitive to local time/longitudinal gradient in the field-line-integrated E-region Pedersen conductivity. The declination angle of the geomagnetic field lines is nearly zero over the Indian and Indonesian sectors. July being the northern summer, the off-equatorial E-region sunset occurs later (earlier) in the northern (southern) magnetic hemisphere with respect to apex sunset over the equator. Therefore, the off-equatorial E-region conductivity in the northern hemisphere should generally be responsible for short-circuiting of dynamo electric fields during July. However, the large enhancement in PRE on this event day (July 28) over both Indian and Indonesian sectors indicates that the short-circuiting effect by the E-region conductivity is absent or kept minimum on this day for some reason, perhaps, due to the forces of lower atmospheric origin.

With a view to further investigate, the mesospheric neutral wind measurements over the Indian sector from medium frequency (2 MHz) partial reflection (PR) radar at an off-equatorial station, Kolhapur, were examined. The geographic location of the Kolhapur PR radar can be seen in Fig. 1. Figure 6 shows the meridional wind variation at 98 km altitude observed from the Kolhapur on July 28 (red curve). The blue curve with error bars represents the 30-day mean variation of meridional wind centered on July 28 (i.e., July 28 ± 15 days) and its standard deviation. The small vertical green arrow on the X-axis indicates the time of equatorial E-region (100 km) sunset over Indian (78° E) sector. It can be noticed from this figure that the meridional wind on July 28 is significantly different from the reference 30-day mean variation and largely southward throughout the day except for a brief interval around midnight. Around sunset (~ 18.5 LT), the meridional wind turns largely southward (equatorward) and reaches approximately − 67 m/s, which is significantly large, compared to its mean value. Further, as will be shown later, this large equatorward wind is a part of quasi-2-day wave that maximizes at 90 km altitude with a vertical wavelength of ~ 107 km. Hence, the meridional wind can be assumed to be in the same phase (direction) over lower thermospheric altitudes of ~ 90–140 km (i.e., 90+ half of the vertical wavelength), however, with a varying magnitude.

The role of thermospheric meridional winds on the strength of PRE and ESF development has been reported in quite a few studies (e.g., Maruyama 1988; Devasia et al. 2002; Jyoti et al. 2004; Abdu et al. 2006a and references therein). Maruyama (1988) has shown that the summer-to-winter transequatorial wind is an important factor controlling the generation of ESF. Assuming a latitudinally uniform transequatorial wind, Maruyama (1988) has shown that the growth rate of RT instability decreases due to the enhanced E-region field-line-integrated conductivity. Later, Devasia et al. (2002) and Jyoti et al. (2004) have suggested that the equatorward meridional wind could help to trigger the instability and ESF development even at lower F-layer heights. Further, Abdu et al. (2006a) have shown that the larger poleward meridional wind can reduce the PRE strength and suppress the ESF development. These studies directly or indirectly show that the equatorward (poleward) meridional wind would enhance (suppress) the PRE and ESF. More recently, Huba and Krall (2013) have shown that the equatorward neutral wind can cause an increase in an upward ExB drift (eastward electric field) at the equator by facilitating the polarization electric fields to grow due to a decrease in off-equatorial E-region conductivity. All the above studies mainly deal with thermospheric meridional winds around F-region altitudes. However, the meridional wind observations around F-region altitudes are not available for the present case. On the other hand, the Kolhapur PR radar observations indicate that the meridional wind is equatorward over lower thermospheric altitudes of ~ 90 to 140 km. In order to examine the role of equatorward/poleward neutral wind over lower thermospheric altitudes on the off-equatorial E-region electron density and Pedersen conductivity, we have employed the SAMI2 (Sami2 is Another Model of the Ionosphere) (Huba et al. 2000) model simulations. Two controlled simulations were carried out with the meridional wind (V) is set to − 60 m/s (equatorward) and + 60 m/s (poleward) over the altitudes 90 to 150 km, and the zonal wind (U) is set to zero for both cases. Figure 7a shows the local time variations of electron density over the off-equatorial location of 10° N dip latitude at 125 km altitude. Figure 7b shows the local time variations of total field-line-integrated Pedersen conductivity over E-region altitudes (90–150 km). From the model simulations presented in Fig. 7a, b, it can be noticed that both E-region density and the field-line-integrated Pedersen conductivity exhibit a small decrease for the case of equatorward meridional wind (blue curves) compared to the case of poleward wind (black curves). Though the decrease in off-equatorial E-region density and field-line-integrated Pedersen conductivity due to the equatorward meridional wind is small, it could have a significant influence on the PRE. For example, Fesen et al. (2000), through TIEGCM model simulations, have shown that a small decrease in E-region electron density from 3300 to 1500 cm⁻³ can cause a significant enhancement in PRE (Fig. 1 of Fesen et al. 2000) as the PRE is highly sensitive to off-equator E-region density around sunset. Hence, one possible explanation could be that the enhanced equatorward neutral wind in lower thermospheric altitudes as observed from Kolhapur PR radar could cause a small decrease in E-region plasma density at off-equatorial latitudes via transport, particularly around sunset hours, where the production due to photoionization is minimum because of higher solar zenith angles. Another possible factor that could cause a significant reduction in E-region density is the divergent node of zonal wind shear via field perpendicular transport of plasma. However, presently, we do not have any observations to support the presence of zonal wind shear. Nevertheless, this decrease in off-equatorial E-region density (or field-line-integrated E-region Pedersen conductivity) could reduce (or remove) the effect of short-circuiting of the F-region dynamo electric fields in the summer hemisphere and that could lead to the enhancement in PRE.

Another mechanism proposed for PRE by Haerendel and Eccles (1992) is the role of equatorial electrojet (EEJ) feeding to the PRE at sunset terminator. Eccles et al. (2015) have shown that the magnitude of PRE is controlled by the longitudinal gradient in the EEJ from the dayside to nightside across the sunset terminator. Haerendel and Eccles (1992) and Eccles et al. (2015) have suggested that the PRE can lift the ionization out from the E-region to further reduce the Cowling conductance and provide a positive feedback for the increase of PRE magnitude. Thus, in order to examine the role of EEJ on the enhanced PRE on July 28, the magnetometer observations from three equatorial stations, Addis Ababa (Africa), Tirunelveli (India), and Phuket (Thailand), that are distributed from dayside to night were examined. The geographic locations of these magnetometer stations can be seen in Fig. 1. For example, Fig. 8a–c show the variations of ∆H (H component of the geomagnetic field after subtracting the quiet day mean midnight value), which is a proxy to the EEJ, from Addis Ababa, Tirunelveli, and Phuket, respectively. The red curves indicate the ∆H variations on July 28, and the blue curves with error bars represent the monthly mean variations with standard deviations. It can be seen from these figures that the ∆H at all the three stations are significantly enhanced on July 28 compared to their respective monthly mean variations. It should be noted that the ∆H variations measured from magnetometer actually represent the integrated current in the eastward direction which is proportional to both conductivity and zonal electric field (J = σ.E). The Hall and Pedersen conductivities maximize at approximately 110 km and 125 km altitudes, respectively (Forbes and Lindzen 1976). Further, the zonal current enhances significantly at equatorial latitudes, called equatorial electrojet (EEJ) due to the Cowling effect. The Cowling conductivity maximizes around ~ 105 km altitude (Forbes and Lindzen 1976). Hence, the major contribution for the ∆H variations measured from the equatorial magnetometers comes from the E-region altitudes due to the enhanced Cowling conductivity in the absence of other solar-driven geomagnetic disturbances (quiet periods). Figure 8d shows the ∆H values from the three stations at 13 UT during which the Addis Ababa is in dayside, Tirunelveli is close to E-region sunset terminator, and Phuket is in the nightside. The longitudinal gradient of ∆H from dayside to nightside on July 28 (red curve) appears to be high compared to its monthly mean variation (blue curve with error bars). This large longitudinal gradient in EEJ on July 28 may also contribute to the enhanced PRE according to the mechanism proposed by Haerendel and Eccles (1992).

Therefore, from the results presented in Figs. 6, 7, and 8, one can envisage that the background conditions such as large equatorward meridional wind at northern off-equatorial E-region (Fig. 6) and enhanced longitudinal gradient of equatorial electrojet from dayside to nightside (Fig. 8) might have caused the favorable conditions for the enhanced PRE and large-scale unseasonal ESF activity on 28 July 2014. The observed changes in the meridional wind and PRE on 28 July 2014 could have a connection with the wave forces of lower atmospheric origin. One of the clues that motivated our line of thinking about the possible role of planetary wave modulation on PRE comes from the daily variation of PSSR during 20–31 July 2014. For example, Fig. 9a, b shows the daily variation of PSSR (peak h′F during 1900–2100 LT) at the stations Bac Lieu and Tirunelveli, respectively. One can readily observe from Fig. 9a, b that the PSSR gradually increases from 20 July to 28 July with a quasi-2-day wave-like modulation at both stations. Further, the PSSR reached maximum on 28 July at both stations.

The results presented in Fig. 9 indicate that a quasi-2-day planetary wave from the lower atmospheric origin might have modulated the PSSR over Indian and Southeast Asian region, possibly, through the modulation in the neutral wind. Abdu et al. (2006b) have shown the planetary wave (~ 2 to 7 day) oscillations in the mesospheric winds can cause similar oscillations in the equatorial evening pre-reversal electric field through E- and F-region electro-dynamical coupling. With a view to further examine this, the meridional neutral wind observations at 98 km altitude from the Kolhapur PR radar were analyzed using a digital bandpass filter. A Butterworth digital bandpass filter centered at 48 h (2 days) with a 3-dB bandwidth of 36–50 h (1.5–2.5 day) is applied on the time series data of hourly meridional wind observations during the months of July to August 2014. Figure 10a shows the filtered quasi-2-day wave oscillations in the meridional wind at 98 km altitude over Indian low-latitude station Kolhapur during July to August 2014. It can be seen from Fig. 10a that the quasi-2-day wave amplitude exhibits significant enhancement during 20–30 July 2014. Further, the amplitude of the quasi-2-day wave increases from 20 July and reaches a maximum around 27–28 July.

Figure 10b shows the amplitude spectra of meridional wind at altitudes from 80 to 98 km at 2 km interval during 20–30 July 2014 computed using least squares method (Malin and Schlapp 1980; Gu et al. 2013). The dominant quasi-2-day wave with a periodicity of ~ 48–52 h can be observed at all altitudes. Figure 10c shows the altitude variations of amplitude and phase of quasi-2-day wave (at 48 h period). The quasi-2-day wave in the meridional wind maximizes at ~ 90 km, and the phase decreases gradually with the height. From the slope of the phase variation, the vertical wavelength of the quasi-2-day wave can be estimated as ~ 107 km. It is known that the vertical wavelengths of quasi-2-day wave are generally large of the order of ~ 70–200 km (e.g., Gu et al. 2013). The large vertical wavelengths of quasi-2-day wave indicate that the meridional wind will be in the same phase (direction) over an altitude range of approximately half of the vertical wavelength. Therefore, it could be possible that the meridional wind would be in the same direction (equatorward) over lower thermospheric altitudes of ~ 90–143 km in the present case. Abdu et al. (2006b) have shown that the neutral wind oscillations associated with the planetary waves can significantly alter the evening PRE via modulating the integrated E-region Pedersen conductivity. Therefore, a quasi-2-day planetary wave in the meridional neutral wind might have caused a similar modulation in PSSR (Fig. 9) through the modulation of the E-region conductivity at off-equatorial latitudes (such as at Kolhapur) and PRE as discussed above (Fig. 6). This study indicates the possible role of planetary wave forces of lower atmospheric origin that can significantly modify the neutral wind at mesosphere and lower thermospheric (MLT) region and PRE system leading to the development of large PSSR and unseasonal large scale EPB/ESF activity. Nevertheless, modeling studies are further needed to substantiate this hypothesis.

An unseasonal enhancement of PRE and development of intense EPBs were observed over Indian and South East Asian longitudes during the post-sunset hours of 28 July 2014. Using multiple independent ground-based observations, the presence of intense EPBs over a wide longitudinal sector ranging from 77° E to 125° E is confirmed. The ATI maps of an EAR for the month of July 2014 indicate that the EPB development on July 28 is an unusual and unique event over this sector. The observations from two equatorial ionosondes at Tirunelveli and Bac Lieu show that F-layer is elevated to higher altitudes prior to the development of these post-sunset EPBs due to the pre-reversal enhancement of zonal electric field. The unusual enhancement of the PRE on this day might be due to the reduction of the E-region conductivity in the northern (summer) low latitudes probably due to enhanced equatorward meridional wind or divergent node of zonal wind shear. Further, it is also speculated that the large longitudinal gradient in equatorial electrojet may also conducive for the enhanced PRE according to the Haerendel and Eccles (1992) mechanism.

A quasi-2-day wave-like oscillation in the PSSR is consistently observed at both equatorial stations, Bac Lieu and Tirunelveli, during 20–31 July 2014. The maximum PSSR occurred on 28 July at both stations. A similar quasi-2-day wave-like oscillation in the meridional neutral wind with a large vertical wavelength of ~ 107 km is observed at the off-equatorial station, Kolhapur, during 20–30 July 2014 with maximum amplitude on 28 July. This indicates the possible coupling between the 2-day planetary wave of lower atmospheric origin and PSSR over Indian and southeast Asian sector. This study emphasizes the coupling from lower atmosphere and the role of planetary wave forcing that can potentially alter the local wind system and E-region conductivities that can lead to the unusual enhancement of PRE and development of EPBs over the wide longitudinal sector, during the season which is climatologically unfavorable for EPBs.