Daytime mid-latitude F2-layer Q-disturbances: A formation mechanism

Negative and positive near noontime prolonged (≥3 hours) F2-layer Q-disturbances with deviations in NmF2 > 35% occurred at Rome have been analyzed using aeronomic parameters inferred from fp180 (plasma frequency at 180 km height) and foF2 observations. Both types of NmF2 perturbations occur under quiet (daily Ap < 15 nT) geomagnetic conditions. Day-to-day atomic oxygen [O] variations at F2-region heights specify the type (positive or negative) of Q-disturbance. The [O] concentration is larger on positive and is less on negative Q-disturbance days compared to reference days. This difference takes place not only on average but for all individual Q-disturbances in question. An additional contribution to Q-disturbances formation is provided by solar EUV day-to-day variations. Negative Q-disturbance days are characterized by lower hmF2 while positive – by larger hmF2 compared to reference days. This is due to larger average Tex and vertical plasma drift W on positive Q-disturbance days, the inverse situation takes place for negative Q-disturbance days. Day-to-day changes in global thermospheric circulation may be considered as a plausible mechanism. The analyzed type of F2-layer Q-disturbances can be explained in the framework of contemporary understanding of the thermosphere-ionosphere interaction based on solar and geomagnetic activity as the main drivers.

Usually F 2 -layer disturbances are related to geomagnetic activity variations but there is a class of F 2 -layer perturbations which occur under quiet geomagnetic conditions (Q-disturbances), their magnitude being comparable to moderate F 2 -layer storm effects. Day-to-day N m F 2 variability is 15-20% at middle latitudes 1,2 while we are speaking about N m F 2 deviations with the magnitude >35% occurring near noontime with a duration more than three hours, i.e. longer than two characteristic e-fold times of the daytime F 2 -layer (the e-fold time is the time required for a parameter to change by 2.72 times). Such long-lasting N m F 2 deviations imply either changes in the ionizing solar EUV radiation or changes in the production/recombination rate (i.e. thermospheric neutral composition) or/and changes in vertical plasma drift mainly related to thermospheric winds. Variations of electric field cannot be excluded but this is less probable during magnetically quiet periods analyzed in the paper.
There is a widely spread opinion that F 2 -layer Q-disturbances are related to the impact from below -the so-called "meteorological control" of the Earth's ionosphere [1][2][3][4][5][6] . However, one has to be cautious to associate all quiet-time F-region disturbances with forcing from the lower atmosphere. Nevertheless accepting that geomagnetic activity is a major cause of F 2 -layer day-to-day variability Rishbeth 2 stressed that the 'meteorological' impact at least was comparable to the geomagnetic one. Despite lots of publications on this topic the actual mechanism of observed day-to-day N m F 2 variations has not been revealed yet. One may find 7,8 that this variability is attributed to 'solar' , 'geomagnetic' and 'other' causes. Therefore any attempt to give a quantitative explanation to day-to-day N m F 2 changes should be considered as one more step towards understanding of the F 2 -layer variability and its prediction, the latter being very important from practical point of view.
The morphology and possible mechanisms of Q-disturbances were earlier discussed [9][10][11][12] . It was shown that both negative and positive daytime Q-disturbances may be related to thermospheric circulation and neutral composition changes. However that time we had not either necessary aeronomic parameters or solar EUV observations to explain the observed Q-disturbances and the number of Millstone Hill ISR observations of Q-disturbances used in previous analyses was very limited. Now using the recently developed method 13 to extract aeronomic parameters from ground-based ionosonde observations, we can reanalyze and specify the formation mechanisms of positive and negative Q-disturbances at least at middle latitudes for daytime hours. Moreover today we have direct solar EUV observations and satellite neutral gas density measurements which can be successfully used to control the obtained results.
The proposed method 13 can be applied when bottom-side Ne(h) profiles are available and such possibility now exists with the worldwide DPS-4 digisonde network 14 . However for various reasons the accuracy of Ne(h) profiles at F 1 -layer heights (used in the method) is different at different stations and not all of them can be used for our aims. There are two ionosondes at Rome (41.9°N; 12.5°E) and DPS-4 observations can be controlled by Italian ionosonde (http://www.eswua.ingv.it/ingv/home.php?res=1024). For this reason Rome DPS-4 Ne(h) observations were used for physical interpretation.
The aims of the paper may be formulated as follows.
(a) to find available cases of negative and positive daytime Q-disturbances using Rome ionosonde observations; (b) to reveal aeronomic parameters responsible for the formation of daytime Q-disturbances in question; (c) to give physical interpretation to the observed daytime Q-disturbances, i.e. to specify their origin.

Method
Selection of the background level is a crucial point dealing with Q-disturbances and various approaches are used to specify it. Our earlier method 9 was based on 27-day running median centered to the day in question. Monthly median or running medians calculated over previous ∼ 30 days are often used as the background [15][16][17] . However any median includes the effects of geomagnetic disturbances occurred during the analyzed period therefore different periods turn out to be in different conditions. Better results should give a selection of magnetically quiet days at a station with binning them in terms of hour, month and range of solar activity. The mean value for each bin provides a quiet-time background level which can be applied with suitable interpolation to any day of a month [18][19][20][21] . But Q-disturbances inevitably contribute to such background level.
Another direction is based on using model monthly median f o F 2 as the background level 22,23 . After averaging (in the model) of many monthly medians obtained under various geomagnetic conditions but similar levels of solar activity one may hope that such average median presents a background level corresponding to a given level of solar activity. Both positive and negative F 2 -layer disturbances (including Q-disturbances) should be seen with respect to such background level.
We may also derive local (for each station) monthly median f o F 2 models 24 which are based on the ionospheric T-index 25,26 as an indicator of solar activity level. It is well-known that effective ionospheric indices of solar activity provide the best correlation with monthly median f o F 2 27 . Such model medians derived for each station are used as the background to calculate f o F 2 deviations. Q-disturbances in our analysis are referred to average over 11, 12, 13 LT hourly (N m F 2 /N m F 2med − 1) × 100% > 35% if all 3-hour ap indices were ≤ the threshold for 24 previous hours. The analyzed negative Q-disturbances took place under very low geomagnetic activity -the average threshold over 24 previous hours on ap was 6.1 ± 2.53 while the average threshold for positive disturbances was slightly higher -ap = 10.1 ± 3.87. A 24-hour preceding time interval was chosen basing on the empirical estimation of the ionosphere reaction to forcing geomagnetic activity. Some estimates of this time constant for mid-latitude F 2 -region are: 12 h 18 , 15 h 28 , 6-12 h 7 ; 16-18 h 16 , 8-20 h depending on season 29 .
The analyzed type of Q-disturbances are not numerous, the larger their magnitude the less they are in number. Our previous morphological analysis 9 has shown that positive Q-disturbances are much numerous compared to negative ones and this is valid especially for noontime perturbations. The selected disturbances should be strong enough to be outside of usual day-to-day 15-20% N m F 2 variability. Further, only the period since 2002 can be analyzed as daily EUV and satellite neutral gas density observations used for a comparison are available for this period. All this taken into account together has specified the magnitude of selected disturbances δN m F 2 = N m F 2 Qday /N m F 2 ref > 35%. Therefore negative disturbances listed in Table 1are all corresponding to the listed requirements which could be found for the analyzed (2002-2018) period.
Besides the background level we need a particular day of the month which would be as much as possible close to the model background N m F 2 . Such days are needed to compare the retrieved aeronomic parameters to those retrieved for the Q-disturbance days. Examples of negative and positive Q-disturbances analyzed in the paper are given in Fig. 1.
If a day is marked as a Q-disturbed one then the sign of N m F 2 deviations (negative or positive) is kept for many hours, sometimes for the whole day as in Fig. 1. This is different from usual geomagnetic storm induced F 2 -layer disturbances when positive and negative phases in N m F 2 variations may change each other in the course of a day. Table 1 gives daytime Q-disturbances found at Rome for the period since 2002. Observed and averaged over (11)(12)(13) LT N m F 2 values are given along with daily F 10.7 and ap(τ) indices 30  Table 1 were developed with the method 13 to retrieve a consistent set of the main aeronomic parameters responsible for the formation of the daytime mid-latitude F-layer. The method 13 is based on solving an inverse problem of aeronomy using observed noontime f o F 2 and five at (10,11,12,13,14) LT values of plasma frequency f p180 at 180 km height as input information along with standard indices of solar (F 10.7 ) and geomagnetic (Ap) activity. Data on f o F 2 and f p180 are available from DPS-4 14 ground-based ionosonde observations after the ionogram reduction. The list of the retrieved parameters includes: neutral composition (O, O 2 , N 2 concentrations), exospheric temperature Tex, the total solar EUV flux with λ ≤ 1050 Å, and vertical plasma drift W mainly related to the meridional thermospheric wind, Vnx. By fitting calculated N m F 2 to observed ones the method provides h m F 2 values which are useful for physical interpretation. For some of the selected cases in Table 1 it was possible to find neutral gas density observations with the Gravity field and steady state Ocean Circulation Explorer (GOCE https://earth.esa.int/web/guest/-/goce-data-access-7219), CHAllenging Minisatellite Payload (CHAMP ftp://anonymous@isdcftp.gfz-potsdam.de/champ/), and Swarm (https://earth.esa.int/web/guest/ swarm/data-access) satellites. For these cases the observed neutral gas density (ρ) was incorporated into our method as a fitted parameter. Available neutral gas density observations in the daytime European sector were reduced to the location of ionosonde and 12 LT using the MSISE00 thermospheic model 31  . Therefore the observed densities were corrected using MSISE00. Depending on conditions this correction may be up to (1-5)%. During this reduction the height of ρ observation was kept unchanged not to introduce an additional uncertainty related to unknown MSISE00 neutral temperature Tex for the particular days in question. The method using observed neutral gas densities was applied when ρ observations were available both for the Q-disturbed and the reference days.  www.nature.com/scientificreports www.nature.com/scientificreports/ Results negative Q-disturbances. Table 2 gives observed δN m F 2 along with the observed total solar EUV (100-1200)Å flux 32 (http://lasp.colorado.edu/lisird/) as well as inferred EUV flux, vertical plasma drift W, F 2 -layer maximum height h m F 2 , atomic oxygen concentration at 300 km in a comparison to MSISE00 model 31 values, and exospheric temperature Tex. Table 2 indicates that selected Q-disturbances are rather strong: δN m F 2 = N m F 2 Qday / N m F 2 ref are (0.5-0.7) with the average value 0.63 ± 0.08. Negative Q-disturbance cases manifest lower F 10.7 and as a rule lower ap(τ) indices compared to the reference days. The average F 10.7 Qday /F 10.7ref = 0.79 ± 0.12 and the average ap(τ) Qday /ap(τ) ref = 0.51 ± 0.36. Lower F 10.7 and ap(τ) indices mean lower ionizing EUV flux and lower level of auroral activity. Table 2 manifests a systematic decrease of atomic oxygen concentration on Q-disturbance days compared to reference days. This is valid not only on average but for particular Q-disturbance cases as well.  Table 2 also indicates that the inferred difference in [O] 300 between Q-disturbance and the reference days is systematically larger than MSISE00 predicts.
Along with decreased [O] concentration on Q-disturbance days ionizing solar EUV flux is also decreased on average by ∼13% due to lower level of solar activity. This decrease is seen both in observed (http://lasp.colorado. edu/lisird/) and retrieved EUV fluxes. Table 2 indicates the average observed EUV Qday /EUV ref ratio 0.88 ± 0.06 while the retrieved average ratio is 0.87 ± 0.08. This closeness of observed and retrieved EUV ratios tells us that the method 13 provides reasonable results.
positive Q-disturbances. The results on positive Q-disturbances are given in Table 3. The selected disturbances are rather strong with δN m F 2 up to 2.2 (the average value =1.66 ± 0.24). In contrast to the negative Q-disturbances F 10.7 and ap(τ) indices on average are larger compared to the reference days: F 10.7 Qday /F 10.7ref = 1.16 ± 0.20 and ap(τ) Qday /ap(τ) ref = 1.78 ± 1.61. This means larger ionizing EUV flux and stronger auroral heating for Q-disturbance days.
In contrast to negative Q-disturbances the concentration of atomic oxygen [O] is systematically larger on Q-disturbance days compared to reference ones. This is valid not only on average but for individual Q-disturbance cases as well.  (Table 3) shows that the model systematically underestimates the atomic oxygen variations. The same situation we met analyzing negative Q-disturbances. Table 3 also demonstrates that Q-disturbance days on average are characterized by larger (∼7%) solar EUV fluxes and this increase is confirmed by direct EUV observations (http://lasp.colorado.edu/lisird/): the inferred average EUV Qday /EUV ref ratio is 1.07 ± 0.09 while the observed ratio is 1.08 ± 0.10. By analogy with negative Q-disturbances ( Table 2) the retrieved and observed average EUV Qday /EUV ref ratios practically coincide and this indicates that the method 13 provides reasonable results as the observed EUV fluxes have nothing common with the retrieval process.
Along with this Table 3 shows some cases with the EUV Qday /EUV ref ratio ≤ 1.0. Electron concentration at F 2 -region heights depends on the O + ion production rate q(O + ) ∼ I EUV [O], where I EUV -the intensity of incident solar EUV radiation. www.nature.com/scientificreports www.nature.com/scientificreports/ it is <1.0 for negative Q-disturbances. The same rule is valid for atomic oxygen concentration as this was shown earlier. Therefore both aeronomic parameters may be used to divide Q-disturbances into negative and positive ones. This result is important from practical point of view as it can be used to predict the type of Q-disturbance.
interpretation. The undertaken analysis has shown that strong (with magnitude >35%) daytime both negative and positive Q-disturbances in N m F 2 are mainly due to day-to-day variations of atomic oxygen concentration.  www.nature.com/scientificreports www.nature.com/scientificreports/ Some but not large contribution to N m F 2 Qday /N m F 2 ref difference provides vertical (downward) plasma drift W presumably related to meridional thermospheric wind Vnx = W/sinIcosI, where I-magnetic inclination at a given location. We are speaking about an effective meridional thermospheric wind. Average W = −17.3 ± 9.1 m/s for negative Q-disturbance days and W = −15.0 ± 8.2 m/s for reference days ( Table 2) while average W = −8.5± 4.3 m/s for positive Q-disturbance days and W = −12.9 ± 4.3 m/s for reference days (Table 3). According to theory of ionospheric F 2 -layer the stronger downward plasma drift the less N m F 2 i.e. stronger poleward Vnx increases the negative Q-disturbance effect while damped poleward Vnx increases the positive Q-disturbance effect.
Negative Q-disturbance days are characterized by lower h m F 2 while positive Q-disturbance days -by larger h m F 2 compared to reference days. Tables 2 and 3 give the average (h m F 2 Qday − h m F 2 ref ) difference of −14. 3 km for negative Q-disturbance cases and this difference is +18.9 km for positive Q-disturbances.
The following expression for h m F 2 obtained from a solution of the continuity equation for electron concentration in the stationary daytime mid-latitude F 2 -layer 35 gives the h m F 2 dependence on aeronomic parameters  (Table 3), the inverse situation takes place for negative Q-disturbance days ( Table 2).
Earlier it was stressed that atomic oxygen day-to-day variations determined the type (positive or negative) of N m F 2 Q-disturbances. Therefore a mechanism of [O] day-to-day variations is a principle question. It is well-known that atomic oxygen is totally produced and lost in the upper atmosphere 36  Day-to-day variations of solar dissociative radiation in Schumann-Runge continuum in principle may contribute to variations of the atomic oxygen abundance when Q-disturbance and the reference days are largely separated in time as the characteristic (e-fold) time of the O 2 dissociation process is ∼ 1/J O2 ≤ 3 days above 120 km height 36 (their Fig. 8.2). However sometimes Q-disturbance and reference days are neighboring ones (e.g. 11/04/2017 and 12/04/2017, Table 3) and this supposes a very fast redistribution of atomic oxygen.
The annual mean eddy diffusion coefficient is ∼4 × 10 6 cm 2 /s at 85-100 km 39 . This gives the characteristic time for the eddy diffusion transfer of atomic oxygen τ tr ∼ H 2 /K edd ∼ 1 day.
This time is comparable to the observed times of atomic oxygen changes during Q-disturbance events. However at present it is not known how eddy diffusion is related to day-to-day changes in solar and geomagnetic activity clearly seen in Q-disturbances occurrence. Moreover K edd manifests seasonal variations 40 which are not followed by Q-disturbance occurrences.
Therefore global thermospheric circulation looks like the most preferable process to explain day-to-day variations of the atomic oxygen abundance. Negative Q-disturbances are associated with extremely low level of magnetic activity corresponding to the minimal intensity of auroral heating. This corresponds to an unconstrained solar-driven thermospheric circulation (a strong poleward neutral Vnx wind during daytime hours) and to relatively low atomic oxygen concentrations at middle latitudes as this follows from the model simulations 41 -low [O] may be related to a moderate upwelling of neutral gas in a wide range of latitudes (their Fig. 3). Indeed, retrieved W for negative Q-disturbance days on average are larger compared to reference days (see earlier) and average [O] Qday /[O] ref ratio = 0.73 ± 0.08. The available neutral gas density observations confirm the obtained results. On negative Q-disturbance days the observed neutral gas density at the satellite height is systematically lower compared to the reference days (Table 2). This difference in neutral gas density is due to different Tex and atomic oxygen concentration.
Similar explanations may be applied to positive Q-disturbance cases. Smaller vertical plasma drifts W on Q-disturbed days ( Table 3) tell us that the northward circulation was damped. This should decrease or even invert upwelling increasing by this way the atomic oxygen abundance in the thermosphere 41,42 .
Direct observations of neutral gas density for Q-disturbance and reference days confirm the obtained results. Four events with ρ observations (Table 3) indicate an increase of neutral gas density on Q-disturbance days. This increase is due to larger neutral temperature and larger atomic oxygen concentration at the analyzed heights  It would be interesting to follow the quantitative contribution of different aeronomic parameters to the observed negative and positive Q-disturbances. This can be done using an approximate expression 43 for N m F 2 in the day-time mid-latitude ionosphere:   (2) is seen to describe satisfactorily the observed δN m F 2 both for the selected negative and positive Q-disturbances ( Table 5). The negative Q-disturbance is due to lower q max and larger β max compared to the reference day while this relation is inversed for the positive Q-disturbance. To a first approximation Table 5 confirms this (see (O/N 2 ) max ratio): 3.24/5.02 = 0.64 (observed 0.59) in the case of negative Q-disturbance and 4.34/2.88 = 1.51 (observed 1.85) for the positive Q-disturbance. The residual difference may be attributed to the difference in EUV. The EUV flux is less on the negative Q-disturbance day and it is larger on the positive Q-disturbance day compared to the reference day (Tables 2 and 3).

Discussion
At the beginning of our investigations we supposed that F 2 -layer Q-disturbances could be related to the impact from below -the so-called "meteorological control" of the Earth's ionosphere [1][2][3][4][5][6] . Seismic events seem also to affect the F 2 -region 44,45 . We ourselves have also considered ionospheric precursors for crustal earthquakes in Italy 46 . However all these morphological and correlation analyses are still at the level of pure speculations concerning physical mechanisms of such impact on the ionosphere. Observed N m F 2 variations with duration longer than e-fold time (∼1.5 hour) imply changes in production and recombination rates i.e. in neutral composition and temperature or changes in thermospheric winds or electric fields. Therefore, telling about the impact from below firstly it is necessary to show which aeronomic parameters responsible for the F 2 -layer formation have been changed and secondly what is the mechanism of such changes, if it is plausible from physical point of view. Unfortunately up to now we don't have necessary observations to answer the formulated questions.
With our recently proposed method 13 the first step in this direction has been done. For a particular geophysical situation using the observed changes of electron concentration in F 1 and F 2 layers the method tells us how the main (controlling) aeronomic parameters have been changed. The method is confined to middle latitudes and noontime hours. However it is sufficient to analyze mid-latitude daytime long-lasting (≥3 hours) Q-disturbances taking place under magnetically quiet conditions.
Using Rome ionosonde observations about forty of analyzed Q-disturbance cases have shown the crucial role of atomic oxygen day-to-day variations in producing F 2 -layer Q-disturbance both negative and positive ones. It was shown that atomic oxygen was systematically decreased on negative Q-disturbance days ( Table 2) and it is systematically increased on positive Q-disturbance days (Table 3) compared to reference days. This result may be considered as a principle one -it was obtained directly from ionospheric observations for the dates of Q-disturbance events. It should be stressed that a good empirical thermospheric model MSISE00 31 does not reproduce (for understandable reasons) the required day-to-day variations of atomic oxygen (Tables 2 and 3).   www.nature.com/scientificreports www.nature.com/scientificreports/ (Tables 2 and 3). All this tells us about the reality of the obtained result on the leading role of atomic oxygen in the formation of Q-disturbances.
Some contribution in a proper direction provides solar ionizing EUV radiation. Both observed and retrieved total solar EUV fluxes indicate on average a decrease for negative EUV Qday /EUV ref = 0.88 ± 0.06 (the retrieved 0.87 ± 0.08) and an increase for positive EUV Qday /EUV ref = 1.08 ± 0.10 (the retrieved 1.07 ± 0.09) Q-disturbances. It should be noted the closeness between the observed and retrieved EUV Qday /EUV ref average ratios. These average results were obtained from individual comparisons given in Fig. 2.
The correlation coefficients between observed and retrieved EUV Qday /EUV ref ratios are significant at the 99.9% confidence level according to Student criterion. This comparison may be considered as an absolutely independent control of the applied method 13 as the observed EUV (http://lasp.colorado.edu/lisird/) fluxes have nothing common with our method to retrieve aeronomic parameters from ionospheric observations. The crucial role of atomic oxygen in the formation of daytime F 2 -layer Q-disturbances has been shown by the undertaken analysis. The other and more difficult question -what is the cause of such day-to-day variations of atomic oxygen? Our earlier discussion of this issue has shown that global thermospheric circulation which is resulted from a competition between solar-driven (background) and storm-induced (due to auroral heating) circulations looks as a plausible mechanism. The inferred W and h m F 2 variations (Tables 2 and 3) clearly indicate changes in the effective meridional wind Vnx when we pass from a reference day to a Q-disturbance one. Under low geomagnetic and lower solar activity compared to the background level we have an unconstrained solar-driven circulation with a strong northward Vnx decreasing the atomic oxygen abundance at F 2 -layer heights 41,42 . These factors along with lower EUV flux work in one direction to decrease N m F 2 and create a negative Q-disturbance. Under slightly elevated geomagnetic and larger solar activity compared to the background level we have damped northward solar-driven circulation. This takes place for two reasons: an elevated auroral heating and an increased ion drag due to larger electron concentration. Damped poleward circulation increases the atomic oxygen abundance in the mid-latitude thermosphere due to downwelling 41 . This along with larger EUV flux creates conditions for a positive Q-disturbance to occur. It should be stressed that all this takes place under quiet geomagnetic conditions when daily Ap <15 nT. Under larger geomagnetic activity (Ap > 15-20 nT) auroral heating increases and inverts the solar-driven thermospheric circulation. The disturbed neutral composition moved by the disturbed equatorward wind spreads from the auroral zone to middle latitudes and we obtain a normal F 2 -layer storm-induced disturbance. Therefore in the case of Q-disturbances we have the same F 2 -layer storm mechanism 41,42,47-49 manifesting itself under low geomagnetic activity.
In this paper we have analyzed the simplest situation with strong Q-disturbances at middle latitudes during daytime hours. According to our morphological analysis 9 such events are not numerous while the majority of Q-disturbances take place during evening and nighttime hours. The nighttime and evening F 2 -layer formation mechanisms are strongly related and meridional thermospheric wind plays an essential role in this mechanism. Future considerations of Q-disturbances in different LT sectors may help clear up the role of thermospheric circulation in day-to-day atomic oxygen variations.
conclusions For the first time negative and positive near noontime prolonged (≥3 hours) F 2 -layer Q-disturbances (daily Ap <15 nT) observed with a ground-based ionosonde at Rome under various seasons and levels of solar activity have been analyzed to reveal their formation mechanism. The recently developed method 13 to extract a consistent set of the main aeronomic parameters from f p180 and f o F 2 observations has been used for this analysis. The obtained results may be summarized as follows.  between Q-disturbance and reference days takes place not only on average but for all individual cases in question. The required atomic oxygen day-to-day variations are not described by the empirical MSISE00 thermospheric model which systematically underestimates the magnitude of these variations. 2. The retrieved atomic oxygen day-to-day variations are confirmed by CHAMP, GOCE, and Swarm satellite neutral gas density observations. Neutral gas density at satellite heights (>400 km) which is mainly presented by atomic oxygen also manifests positive and negative deviations in accordance with the observed type of Q-disturbance. 3. An additional contribution to Q-disturbances formation is provided by solar EUV day-to-day variations.
Both observed and retrieved total solar EUV fluxes indicate on average a decrease for negative EUV Qday / EUV ref = 0.88 ± 0.06 (the retrieved 0.87 ± 0.08) and an increase for positive EUV Qday /EUV ref = 1.076 ± 0.10 (the retrieved 1.074 ± 0.09) Q-disturbances. The closeness between observed and retrieved average EUV Qday /EUV ref ratios is a direct confirmation for the reality of the obtained results as the observed EUV fluxes have nothing common with the retrieval process. 4. Negative Q-disturbance days are characterized by lower h m F 2 (on average −14.3 km) while positive Q-disturbance days -by larger h m F 2 (on average +18.9 km) compared to reference days. This is due to larger average Tex and vertical plasma drift W on positive Q-disturbance days compared to reference ones, the inverse situation takes place for negative Q-disturbance days. 5. Day-to-day changes in global thermospheric circulation manifested by W day-to-day variations are considered as a plausible mechanism. The analyzed type of F 2 -layer Q-disturbances can be explained in the framework of contemporary understanding of the thermosphere-ionosphere interaction based on solar and geomagnetic activity as the main drivers.