The Christmas Island VHF radar provides data of meter-scale F-region irregularities routinely, initially operated by the Stanford Research Institute – SRI International (2002–2007) – and then operated by the US Air Force Research Laboratory (AFRL). The system uses a 100 m × 100 m coaxial collinear (COCO) antenna array. Two stationary beams were used for measurements. One beam is pointed north (azimuth 0∘ and elevation 84.5∘), and the other one is pointed to the east (azimuth 90∘ and elevation 60.5∘). The coherent radar detects fluctuations related to the plasma instabilities called field-aligned irregularities, and then detection of such irregularities requires the antenna to be pointing perpendicularly to the geomagnetic field line (Tsunoda et al., 1979, 2000). Then, the northern beam antenna was chosen due to be pointed in the northerly direction to reach perpendicularly to the magnetic field line. More technical details of this radar can be found in Miller et al. (2009). Its geographic position is 2.0∘ N, 157.4∘ W, 2.9∘ N dip latitude, and its magnetic inclination (declination) varied from 4.69∘ (9.36∘ E) in 2003 to 4.61∘ (9.38∘ E) in 2012.

It is worth mentioning that measurements available to this study cover different solar conditions when F10.7 varied from 200 SFU (high solar flux conditions) to 66 SFU (low solar flux conditions), as shown in Fig. 1. Data measurements of spread-F echoes to this study are in the periods of January 2003 and December 2012. All our data are presented as altitude integration from 200 to 1000 km height as a function of the signal-to-noise ratio (present in Fig. 2) and the horizontal dashed lines (at 20:00 and 00:00 LT) representing a time threshold to assist observations of time–echo distribution. Lack of data is also presented as a black space in the figure, mainly for 2014 (March equinox and June solstice).

Our interest focus is the local occurrence of F-region echoes (5 m scale irregularities) as one of the most interesting and challenging phenomena for space weather and climatological models. The physical mechanisms responsible for this phenomenon are complex and not fully understood. So, we organized our data attempting to present the difference in seasonal and solar flux conditions as a function of the time and height of irregularities observed in the VHF radar. For this study, we limit our focus to quiet-time irregularities.

It is well known that high geomagnetic activities directly cause drastic perturbations in the zonal electric field and in the equatorial and low-latitude regions, affecting the growth and development of ionospheric irregularities. These perturbations can be categorized as prompt penetration (PP) and disturbance dynamo (DD) electric fields (Abdu et al., 2018; Astafyeva et al., 2018; Shreedevi and Choudhary, 2017). These perturbed electric fields occurring in the post-sunset period can enhance/weaken the regular eastward electric field and vertical plasma drift and then affect the uplift of the F layer (Fejer et al., 1991), as a consequence affecting the generation of irregularities (Aarons, 1991; Abdu, 2012).

In sequence, to avoid the disturbed geomagnetic periods and their effects on irregularity generations, we classify the data with low geomagnetic conditions using the 3 h planetary K index (known as Kp). Each measurement was tagged with the value of Kp for the time of the measurement plus the previous 3 Kp values. We limited our study to quiet geomagnetic conditions to be those when none of the three Kp indexes exceeded 3.

The solstice is when the Sun reaches the most southerly or northerly points in the sky, while an equinox is when the Sun passes over Earth's Equator. So, to sort our measurements according to the four seasons spring, summer, fall, and winter, we use 91 d of data centered on each day 21 of March, June, September, and December, respectively. We used the quiet-time radar echoes for each season to obtain the occurrence rate of echoes. We establish that a good representation of irregularity occurrence is given by the echo distribution above 0 dB divided by the total number of observations. Our criterion is a good commitment between being able to identify the occurrence of spread-F echoes and to eliminate the effects of non-geophysical echoes.

The sample rate of the VHF radar is estimated for every 15 min interval starting at 18:00 LT, right before sunset until 05:00 LT near sunrise. To construct maps of the irregularity occurrence rate as a function of height and local time, we had computed for every 15 km height intervals starting at 200 km up to 1000 km altitude.

Peak altitude profiles of the occurrence rate of F-region echoes are shown in Fig. 3, top panels. They were organized by season (March equinox, June solstice, September equinox, and December solstice, from left to right, respectively) along the 2003 to 2012 period. Horizontal dashed lines are placed at the 250 and 350 km heights to assist observation (hereafter called the altitude threshold).

Comparing the peak of altitude echoes along seasons, we can observe higher occurrence rates of all years over the June solstice and September equinox than the March equinox and December solstice seasons. This observation matches the previous result by Cueva et al. (2013) that showed the peak occurrence of equatorial spread F for this region being around the July–August months.

When analyzing solar minimum years (2006 and 2008), we can focus our attention on the peak echo altitude, which was slightly higher, in altitude, in the June solstice than in the September equinox. As observed in Fig. 3, we found a more prominent peak time/altitude occurrence in the September equinox (before midnight) than in the June solstice (around midnight hours). For years of solar maximum (2003 and 2012), we can mention that peak altitude distribution is highest, mainly during the June and September seasons, but nevertheless presents a minor percentage of occurrence than solar minimum years (2006 to 2008). The minimum occurrence of peak altitude occurs in the March equinox, which is the period of scarce spread-F echoes over the Christmas Island region. During the solar maximum period spread-F echoes have less occurrence than in the solar minimum period, reaching higher altitudes as observed in the 2003 June solstice, when the peak altitude was higher than the threshold altitude of 350 km. During the September equinox higher plumes are frequently observed than in other periods, which agrees with results presented by Cueva et al. (2013).

Time variation in the occurrence rates of F-region echoes for the period of the study is shown in Fig. 3, bottom panels, also separated by seasons (March equinox, June solstice, September equinox, and December solstice, from left to right, respectively). The vertical dashed lines represent local sunset and local midnight. As we can observe, the percentage of occurrence of echoes presents solar flux dependence. During solar maximum radar echoes are confined to a few hours after sunset; on the other hand, during solar minimum echoes are more broadened out in time and can arise late in the evening after sunset and more closely to midnight hours. As we get closer to the solar minimum period, the amplitude of echo occurrence increases due to the high probability of echoes occurring all through the night, which is observed during the years 2006 to 2008 with more amplitude than echoes observed during solar maximum. A similar finding was mentioned by Niranjan et al. (2003) when analyzing spread-F data from the 1997–2000 period and also by Burke et al. (2004) and Dao et al. (2011) using satellite data from different geographical locations.

Seasonal dependence of echoes along the solar cycle is also observed. The September equinox has more conditions to develop irregularities over the region, as explained before, as well as higher echo occurrence for either the solar minimum or maximum periods. For the March equinox and December solstice, we have less probability of echo occurrence as observed in Fig. 3; moreover, the amplitude of echo occurrence is always lower in the solar maximum than in the solar minimum. It is important to mention that, for the June solstice, the echoes are especially observed around local midnight and post-midnight hours, which is in agreement with observations made by Otsuka et al. (2018) during the solar minimum period.

Under quiet magnetic conditions and solar minimum, there are some possible seeding mechanisms competing that increase the probability of spread-F generation throughout the night (pre-midnight and post-midnight) as well as uplifting of the F layer. For example, gravity waves, launched from the active convection region in the troposphere, could propagate into the ionosphere (Takahashi et al., 2009, 2010; Maurya et al., 2020; Correia et al., 2020) and contribute to the instability seeding. Another is the medium-scale traveling ionospheric disturbance (MSTID) activity providing perturbations in the electric fields for the low-latitude F region to be unstable at post-midnight hours, which can seed the RT instability at the magnetic equator (Otsuka et al., 2009; Yokoyama et al., 2011; Narayanan et al., 2019). Another mechanism could be the uplift of the F layer around midnight (Nicolls et al., 2006) caused by a decreasing westward electric field in conjunction with sufficient recombination and plasma flux. However, the causes of midnight F-layer increase are not yet clearly established.

For the extended solar minimum period, during the June solstice and December solstice months, we observed post-midnight echoes similar to those previously reported by Otsuka et al. (2012) during the 2005 to 2009 period. The September equinox also presents post-midnight events for the solar minimum period. Our findings are summarized in Fig. 4. In panel a is presented UT (LT = UT + 14) on the vertical axis for the time peak echo occurrence along the solar cycle separated by seasonality. The ionospheric sunset and the local midnight are highlighted as horizontal dashed lines, and the error bars represent the standard deviations. We can clearly observe that the peak time echo occurrences are closer to the time of the PRE during high solar activity years, with small standard deviations (see 2003, 2004, 2011, and 2012) and around midnight during solar minimum conditions, with bigger standard deviations (see years 2007 to 2009). The December solstice season shows very different behavior in the years 2004 and 2012 during the high solar condition; also in 2003, the peak time echo occurrence was very late compared with the other seasons for the same year. So, further study must be necessary at this point.

According to our observations, during solar minimum, the error bars (standard deviation) must be higher than during solar maximum periods due to the probability of spread-F echo occurrences, which are spread out in time (with maximum observations around local midnight) under minimum conditions and localized to around local sunset under maximum solar conditions. For example, the March equinox and June solstice are good representations of this behavior. They show very small deviation bars in 2003, 2011, and 2012 (solar maximum conditions) and large deviation bars from 2006 to 2009 (solar maximum conditions). The September equinox represents the same general trend, but during 2012 the deviation bar is large, beyond that expected for solar maximum conditions.

Figure 4b shows altitude peak variation along the solar cycle, also separated by seasonality. The error bars show the distribution of peak altitude echo observations. The altitude parameter seems to follow a very good trend, with higher altitudes for solar maximum conditions and lower altitudes for solar minimum conditions. Again, the December solstice does not match very well with this trend, also presenting a lower altitude peak compared with all seasons. During 2008 the December solstice is higher than its general trend. We observe the size of error bars decreasing from solar maximum to solar minimum conditions. During solar maximum, the echoes reach higher altitudes compared with echo occurrences during solar minimum, which is why the deviation bars are bigger during solar maximum years. The altitude parameter is an important parameter since it is one key process in the generation mechanism for ionospheric irregularities. Peak altitude echoes of the June solstice reach a higher altitude difference from solar maximum to solar minimum periods when compared with the March and September equinoxes, which were closer to 300 km for most of the solar cycle period.