Solar zenith angle dependencies of F 1-layer , Nm F 2 negative disturbance , and G-condition occurrence probabilities

Experimental data acquired by the Ionospheric Digital Database of the National Geophysical Data Center, Boulder, Colorado, from 1957 to 1990, are used to study the dependence of the G condition, F1-layer, and NmF2 negative disturbance occurrence probabilities on the solar zenith angle during summer, winter, spring, and autumn months in latitude range 1 (between −10 and+10 of the geomagnetic latitude,8), in latitude range 2 (10 ◦ < |8| ≤ 30), in latitude range 3 (30 ◦ < |φ| ≤ 45, 30 < |8| ≤ 45), in latitude range 4 (45 ◦ < |φ| ≤ 60, 45 < |8| ≤ 60), and in latitude range 5 (60 ◦ < |8| ≤ 90), whereφ is the geographic latitude. Our calculations show that the G condition is more likely to occur during the first half of a day than during the second half of a day, at all latitudes during all seasons for the same value of the solar zenith angle. The F1-layer occurrence probability is larger in the first half of a day in comparison with that in the second half of a day for the same value of the solar zenith angle in latitude range 1 for all seasons, while the F1-layer occurrence probability is approximately the same for the same solar zenith angle before and after noon in latitude ranges 4 and 5. The F1-layer and G condition are more commonly formed near midday than close to post sunrise or pre-sunset. The chance that the daytime F1-layer and G condition will be formed is greater in summer than in winter at the given solar zenith angle in latitude ranges 2–5, while the F1-layer occurrence probability is greater in winter than in summer for any solar zenith angle in latitude range 1. The calculated occurrence probability of the NmF2 weak negative disturbances reaches its maximum and minimum values during daytime and night-time conditions, respectively, and the average night-time value of this probability is less than that by day for all seasons in all studied latitude regions. It is shown that the NmF2 normal, strong, and very strong negative disturbances are more frequent on average at night than by day in latitude ranges 1 and 2 for all seasons, reaching their maximum and minimum occurrence probability values at night and by day, respecCorrespondence to: A. V. Pavlov (pavlov@izmiran.rssi.ru) tively. This conclusion is also correct for all other studied latitude regions during winter months, except for the NmF2 normal and strong negative disturbances in latitude range 5. A difference in the dependence of the strong and very strong NmF2 negative disturbance percentage occurrences on the solar zenith angle is found between latitude ranges 1 and 2. Our results provide evidence that the daytime dependence of the G condition occurrence probability on the solar zenith angle is determined mainly by the dependence of the F1-layer occurrence probability on the solar zenith angle in the studied latitude regions for winter months, in latitude range 2 for all seasons, and in latitude ranges 4 and 5 for spring, summer, and autumn months. The solar zenith angle trend in the probability of the G condition occurrence in latitude range 3 arises in the main from the solar zenith angle trend in the F1-layer occurrence probability. The solar zenith angle trend in the probabilities of strong and very strong NmF2 negative disturbances counteracts the identified solar zenith angle trend in the probability of the G condition occurrence.


Introduction
The Ionospheric Digital Database of the National Geophysical Data Center, Boulder, Colorado, provides the routine sounding ground-based station measurements of the critical frequencies and virtual heights of different ionospheric layers, and, in particular, the critical frequencies fof1 and fof2 of F1-and F2-layers that are analyzed in this study.The values of the peak densities, NmF1 and NmF2, of the F1-and F2-layers are related to the critical frequencies fof2 and fof1 as NmF2= 1.24 × 10 10 fof2 2 and NmF1= 1.24 × 10 10 fof1 2 , where the unit of NmF2 and NmF1 is m −3 , the unit of fof2 and fof1 is MHz (URSI handbook of ionogram interpretation and reduction, 1978).The Ionospheric Digital Database is formed using the URSI standard rules (URSI handbook of ionogram interpretation and reduction, 1978).In addition to numerical values of ionospheric parameters, the qualifying and descriptive letters A-Z are used in this database.The descriptive letter G means that a measurement is influenced by, or impossible because, the ionization density of the layer is too small to enable it to be made accurately, and this case is described as a G condition in the F-region of the ionosphere when fof2 ≤ fof1 (URSI handbook of ionogram interpretation and reduction, 1978).If the layer is not seen from ionograms due to other reasons, then other letters are used.The G condition arises in the ionosphere when the critical frequency of the F2-layer drops below that of the F1-layer, i.e. when the peak density, NmF1, of the F1-layer, which is composed mostly of the molecular ions NO + and O2 + , is larger than that of the F2-layer, which is dominated by O + ions (King, 1962).As a result, a very low main peak altitude value (below 200 km) is observed in ionograms, so that no information is obtainable above this height from ground-based ionozonde data.As far as the authors know, the first altitude distribution of the electron density during a G condition was deduced by Norton (1969) from ionograms recorded by the Alouette I satellite ionozonde and the St. John's ground-based ionozonde during the severe negative ionospheric storm on 18 April 1965.
The physics of the G condition phenomenon has been studied by Buonsanto (1990) using ionosonde data from two mid-latitude stations, by Oliver (1990), using Millstone Hill incoherent scatter radar data, and by Fukao et al. (1991) using data from the middle and upper atmosphere radar in Japan.Pavlov and Buonsanto (1998), Pavlov (1998), Pavlov et al. (1999), Schlesier and Buonsanto (1999), and Pavlov and Foster (2001) studied the G condition formation for quiet and disturbed mid-latitude ionosphere during periods of low, moderate, and high solar activity, using the Millstone Hill incoherent scatter radar data.Model results also show that O + can become a minor ion in the F-region, creating a G condition during disturbed conditions at high latitudes (Banks et al., 1974;Schunk et al., 1975), and observations at EIS-CAT confirm this conclusion (e.g.Häggström and Collis, 1990).These papers provide evidence that changes in [O], [N 2 ], [O 2 ], and the plasma drift velocity, the effect of the perpendicular (with respect to the geomagnetic field) component of the electric field on the electron density (through changes in the rate coefficients of chemical reactions of ions), and the effects of vibrationally excited N 2 and O 2 on the electron density are important factors that control the G condition formation in the ionosphere.The study of the G condition formation in the ionosphere above Millstone Hill during the severe geomagnetic storm of 15-16 July 2000, provided a weighty argument for the inclusion of the effects of vibrationally excited N 2 and O 2 on the electron density and temperature in ionospheric models (Pavlov and Foster, 2001).
During NmF2 disturbances, which are believed to be caused by geomagnetic storms and substorms, the value of NmF2 can either increase or decrease in comparison with a geomagnetically quiet NmF2, and these changes are denoted as positive and negative disturbances (Prölss, 1995;Buonsanto, 1999).A decrease in NmF2 during a NmF2 negative disturbance leads to an increase in the G condition occurrence probability if the F1-layer exists.On the other hand, the G condition cannot exist in the ionosphere if there is no F1-layer.The preceding work by Lobzin and Pavlov (2002) summarizes papers addressing the measurements and the physics of the F1-layer, NmF2 negative disturbance, and G condition, and gives for the first time the detailed dependencies of the probabilities of the F1-layer, NmF2 negative disturbance, and G condition occurrences on a daily solar activity index, F10.7, a 3-h geomagnetic index, K p , a number of a given day in a year, and a geomagnetic latitude.The aim of this paper is to carry out a statistical study of solar zenith angle dependencies of NmF1, NmF2 negative disturbance, and G condition occurrence probabilities using the Digital Database fof1 and fof2 measurements from 1957 to 1990.
Some features of the solar zenith angle dependencies of NmF1, NmF2 negative disturbance, and G condition occurrence frequencies have been known for a long time.Ratcliffe (1956Ratcliffe ( , 1972)), Yonezawa et al. (1959) concluded that the F1layer is less liable to appear for larger values of the solar zenith angle.Du Charme and Petrie (1973) derived an expression to predict fof1, assuming limits for the presence of the F1-layer as a function of the solar zenith angle and of solar activity.Scotto et al. (1997) tested the Du Charme and Petrie (1973) formula adopted in the International Reference Ionosphere (IRI) model, taking into account alternative solutions for the particular restrictions imposed by the IRI for high values of solar zenith angle.New probability functions to predict the occurrence of the F1-layer have been proposed by Scotto et al. (1997Scotto et al. ( , 1998) ) to replace the Du Charme and Petrie (1973) formula.
A negative F2 ionospheric storm onset at middle latitudes is most frequently observed in the morning time sector and very rarely in the noon, afternoon, and night-time sectors (Prölss, 1995;Buonsanto, 1999).Wrenn et al. (1987) discriminated geomagnetic activity levels as very quiet, quiet, normal, disturbed, and very disturbed conditions.The negative ionospheric storm effect in NmF2 during normal, disturbed, and very disturbed conditions is centered at night for very disturbed conditions and during morning hours for normal and disturbed conditions if the ionozonde fof2 measurements from the Argentine Islands ionozonde station are used (Wrenn et al., 1987).The comparison between the summer fof2 measurements of the Argentine Islands and Port Stanley ionozonde stations leads to the conclusion that the maximum fof2 depression moves from the night-time sector to the morning sector if the latitude of the station is changed from middle to more low latitudes (Wrenn et al., 1987).It is found by Ratcliffe (1972) that the G condition is more commonly formed near midday than during several hours after sunrise or before sunset.However, the results of Ratcliffe (1972) are not formulated in a mathematical form to be used in calculations.
Previous investigations are based on the limited ionozonde data set from some stations and on theoretical analysis of the main physical processes that form electron density altitude profiles.As a result, except for the Scotto et al. (1997Scotto et al. ( , 1998) ) formulas for the F1-layer probability function, there are no published solar zenith angle dependencies of NmF1, NmF2 negative disturbance, and G condition occurrence probabilities.The main purpose of this work is to calculate for the first time these probabilities for low, middle, and high latitudes in summer, in winter, and during the spring and autumn months, to provide some quantitative measure of these probability variations.In our analysis we study for the first time a possible relationship between the solar zenith angle probability dependence of the G condition occurrence with the solar zenith angle probability dependencies of NmF2 negative disturbance and F1-layer occurrences.
2 Formation of the F1-and F2-layers in the ionosphere Solar zenith angle dependencies of NmF1 and NmF2 negative disturbance and G condition occurrence probabilities, which are studied in our work, are determined by physical processes that form the F-region of the ionosphere.The F-region is located in the altitude range above 140-160 km.Within the F-region are the F1-and F2-layers, with the peak altitudes hmF1< 190 − 200 km and hmF2> 200 − 210 km, respectively.The major F1-and F2-layer ions are O + ( 4 S), O + 2 , and NO + .The main physical processes that form the F1 and F2-layers in the ionosphere by a balance between production, chemical loss, and transport of electrons and ions are described in many review articles, books, and papers (e.g.Ratcliffe, 1972;Rishbeth and Garriot, 1969;Brunelli and Namgaladze, 1988;Rees, 1989;Fejer, 1997;Rishbeth and Muller-Wodarg, 1999;Rishbeth, 2000;Rishbeth et al., 2000;Abdu, 2001;Lobzin and Pavlov, 2002;Pavlov and Foster, 2001).Following these studies, it is usually supposed that the value of NmF2 is approximately directly proportional to the [O]/L ratio at hmF2 during daytime conditions, where L is the loss rate of O + ( 4 S) ions in the reactions of O + ( 4 S) with unexcited N 2 (v = 0) and O 2 (v = 0) and vibrationally excited N 2 (v) and O 2 (v) molecules at vibrational levels, v > 0. Thus, the depletion in [O] and the increase in [N 2 ] and [O 2 ] can lead to a negative phase in NmF2.The increase in the rate coefficients for reactions between O + ( 4 S) ions and N 2 (v ≥ 0) and O 2 (v ≥ 0), due to changes in neutral and ion temperatures and due to the increase in vibrational temperatures of N 2 and O 2 would also produce negative storm effects in NmF2.These assumptions are used in our study in discussions of NmF2 variation sources and to understand reasons for solar zenith angle dependencies of NmF2 negative disturbance and G condition occurrence probabilities.
To illustrate the basic physics involved and to study the physical reasons for NmF2 negative disturbance occurrence probability nighttime variations, it is useful to use the analytical approximation of the nighttime mid-latitude NmF2 given by Krinberg andTashchilin (1982, 1984) as where t is a local time, NmF2(t 0 ) is the ionospheric electron density for the local time t 0 corresponding to dusk, T n is the exospheric neutral temperature, T i and T e are ion and electron temperatures, g is the acceleration due to gravity, ν in is the O + −O collision frequency, F ∞ is the value of plasma ion flux flowing from the plasmasphere into the ionosphere, the values of L, T i , T e , g, and ν in are chosen at the F2 peak altitude.
One can see from Eq. ( 1) that the nighttime F2-region electron density consists of two parts.The first term describes the role of the daytime ionization in the maintenance of the nighttime ionosphere.In this case the F-region would decay with the characteristic time ∼ L −1 (about several hours).Since the loss rate of O + ( 4 S) ions is proportional to [N 2 ] and [O 2 ], an increase or decrease in [N 2 ] and [O 2 ] at hmF2 altitudes leads to a decrease or an increase in NmF2, respectively.The nocturnal F-region is also maintained by a downward flow of ionization from the plasmasphere, described by the second term in Eq. (1).In winter, and possibly in spring and in autumn, when the night is long enough, the role of the second term in Eq. (1) increases before sunrise, and this term can determine the mid-latitude value of NmF2.
The role of the ion transport is less than the role of the chemical reactions of ions with electrons and neutral components of the upper atmosphere at the F1-layer altitudes, and the production and loss rates of electrons and ions that determine the F1-layer formation.To study the formation of the F1-layer, Ratcliffe (1972) assumed that the main source of NO + ions is the chemical reaction of O + with N 2 , and there are only NO + and O + ions.Ratcliffe (1972) found that the peak of the F1-layer exists in the ionosphere if the peak altitude, h 0 , of the total production rate of thermal electrons is less than an altitude, h t .The value of h t is determined from the condition of K [N 2 ] = α[e], where K is the rate coefficient for the reaction of O + ions with N 2 , and α is the rate coefficient of the dissociative recombination of NO + ions.Ratcliffe (1972) concluded that the value of h t − h 0 is decreased with the solar activity level increase, and the value of h t − h 0 has a maximum value close to midday.As a result, the F1 peak is more clearly in evidence at solar minimum than at solar maximum, and the F1 peak is more commonly formed near midday and in summer (Ratcliffe, 1972).Yonezawa et al. (1959) carried out another simple consideration for the F1-layer to appear as a distinct layer.In addition to the equality of α and the rate coefficient of the dissociative recombination of O + 2 ions, the height gradient of the atmospheric neutral components, the rate coefficients of O + ions with N 2 and O 2 , and the value of α were assumed to be constants by Yonezawa et al. (1959), to obtain the condition of appearance of the F1-layer as where Q 0 is the maximum production rate of electron-ion pairs for the Sun by photoionization of atomic oxygen at an altitude, h = h 0 .
Using Eq. ( 2), Yonezawa et al. (1959) concluded that the F1-layer is more liable to appear during periods of low solar activity than during periods of high solar activity, and during the day near noon than near sunrise or sunset.It follows from Eq. (2) that the NmF1 occurrence probability approaches 100% if Eq. ( 2) is valid and 0% if Eq. ( 2) is not realized in the ionosphere.
The ionosphere at low latitudes is very sensitive to electric fields.The daytime low latitude electric field that is directed eastward causes the ionosphere to be lifted to high altitudes along magnetic field lines, where there is a very rapid diffusion of electrons and ions, and gravity pull the electrons and ions downward and poleward on either side of the magnetic equator, so that a low latitude trough develops over the equator, with the F2-layer density maximum to the north and south (Rishbeth, 2000;Abdu, 2001).This F2-layer structure is usually called the equatorial or Appleton anomaly.If the daytime eastward electric field is strengthened or weakened during geomagnetic storms and substorms, the F2-layer density maximum move further poleward or equatorward, and the ionospheric density over the equator is reduced or increased, respectively.It is remarkable that the upward drift by day is balanced by a downward drift at night, and hmF2 is lower at night than by day at low latitudes, while the middle latitude hmF2 is higher at night than by day.According to Sterling et al. (1972), the low-latitude F1-layer is caused by the electromagnetic drift, rather than by an effect of F1-layer photochemistry.

Data and method of data analysis
Ionograms produced by ionozondes are records that show variations of the virtual height of radio wave reflection from the ionosphere as a function of the radio frequency, h (f ), within the frequency band range 1 MHz-20 MHz that is normally used (URSI handbook of ionogram interpretation and reduction, 1978).The radio wave that is reflected from the ionosphere level of ionization is split into two waves of different polarization in the Earth's magnetic field, thereby leading to two sorts of observed h (f ) curves.These waves are called the ordinary wave (o-mode) and the extraordinary wave (x-mode).There are also z-mode traces on some ionograms generated by radio waves which have been propagated along the magnetic field lines.The mode traces can be identified by the frequency separation and by other indications presented in the URSI handbook of ionogram interpretation and reduction (1978).A simple approach is used to find peak electron densities of the ionosphere from observations of h (f ) curves.When the level of the peak electron density in the layer is reached, the value of h (f ) becomes effectively infinite ( df dh → 0).The frequency at which this occurs is determined as the critical frequency of the ionospheric layer.
Our analysis is based on 34 years of hourly fof2 and fof1 data from 1957 to 1990 from stations available on the Ionospheric Digital Database of the National Geophysical Data Center, Boulder, Colorado.At the chosen ionozonde sta-tion, the solar zenith angle, χ , is a function of a local time, a geographic latitude, and a number, n d , of a given day in a year.Therefore, a multiple-parameter statistics is needed to study the solar zenith angle dependencies of NmF2 negative disturbance, NmF1, and G condition occurrences.We do not analyze the ionozonde measurements of fof2 and fof1 in the Northern Hemisphere and the Southern Hemisphere separately, but carry out our statistical analysis of solar zenith angle dependencies of NmF1, NmF2 negative disturbance, and G condition occurrences separately in summer (June, July, and August in the Northern Hemisphere, and December, January, and February in the Southern Hemisphere), in winter (December, January, and February in the Northern Hemisphere, and June, July, and August in the Southern Hemisphere), and during spring and autumn months (March, April, May, September, October, and November in both hemispheres).
As we have pointed out, the solar zenith angle is a function of geographic latitude, ϕ.Therefore, to study the solar zenith angle dependencies of NmF2 negative disturbance, F1-layer, and G condition occurrences, the geographic latitude range has to be taken so that this range is minimized, while the number of measurements remains large enough to carry out this statistical study.On the other hand, there are significant differences in physical processes that determine NmF2 negative disturbance, NmF1, and G condition occurrences at low, middle, and high geomagnetic latitudes, (e.g.Ratcliffe, 1972;Rishbeth and Garriot, 1969;Brunelli and Namgaladze, 1988;Rees, 1989;Fejer, 1997;Rishbeth and Muller-Wodarg, 1999;Rishbeth et al., 2000;Lobzin and Pavlov, 2002), and these differences can lead to differences in the solar zenith angle dependencies of the studied events.Therefore, we split the ionozonde fof2 and fof1 data set used into five parts.A geomagnetic equatorial region (−10 • ≤ ≤ 10 • ), where an equatorial daytime NmF2 trough is developed, is defined in our study as a latitude range 1.A latitude range 2 is a low-latitude region (10 • < | | ≤ 30 • ) where daytime NmF2 crests in comparison when equatorial daytime NmF2 are observed.Latitude ranges 3 and 4 are mid-latitude regions.A latitude range 3 includes both 30 It is clear from the definition of these mid-latitude ranges that latitude ranges 3 and 4 are not overlapping.
The main ionization trough, the ionization hole in the polar cap around local dawn, the tongue of ionization, and the aurorally produced ionization electron density peak in the vicinity of the auroral oval are a natural consequence of the difference and competition between the various chemical and transport processes known to be operating in the high-latitude F-region ionosphere (for more details, see, e.g.Rees, 1989;Buonsanto, 1999).The latitude and longitude boundaries of these regions show marked variations.As a result, only average solar zenith angle dependencies of NmF1, NmF2 negative disturbance, and G condition probability functions in a latitude range 5 (60 are calculated in this paper.To discriminate between the morning and evening solar zenith angle dependencies of the studied events, we split the range of 0 • ≤ χ ≤ 180 • into twelve intervals of the same length, χ , both from 00:00 LT to 12:00 LT, and from 12:00 LT to 24:00 LT at each ionozonde station. We consider the measured fof1 and fof2 within the abovedetermined local time, latitude, and month range, and determine the probability, G (χ ), or F 1 (χ ), of the G condition or F1-layer occurrence as a ratio of the number of G condition or F1-layer observations for zenith angles within the χ interval to a total number of measurements for the same χ and within the given local time, latitude, and month range.
The electron density can either decrease or increase during geomagnetically disturbed conditions, and these electron density changes are denoted as negative and positive ionospheric disturbances, respectively.To test the effects of geomagnetic activity, we use two different K p labels: "disturbed", for which we take K p > 3 and use the peak density, NmF2(d), and critical frequency, fof2(d), of the F2-layer observed during the time periods with K p > 3, and "quiet", for which we take K p ≤ 3. The determination of the quiet peak density, NmF2(q), and critical frequency, fof2(q), of the F2-layer, is crucial for studies of negative and positive ionospheric disturbances.
Perturbations in the neutral composition, temperature, and wind at one altitude are rapidly transmitted to higher and lower altitudes.However, it takes time to relax back to an initial state of the thermosphere, and this thermosphere relaxation determines the time for the disturbed ionosphere to relax back to the quiet state.It means that not every fof2 observed during the day with K p ≤ 3 can be considered as fof2(q).The characteristic time of the neutral composition recovery after a storm impulse event ranges from 7 to 12 h on average (Hedin, 1987), while it may need up to several days for all altitudes down to 120 km in the atmosphere to recover completely back to the undisturbed state of the atmosphere (Richmond and Lu, 2000).As a result of this thermosphere recovery, a day with K p ≤ 3 from 00:00 UT to 24:00 UT cannot be considered a quiet day if the previous day was a day with K p > 3 from 00:00 UT to 24:00 UT.We determine the quiet reference day with fof2(q) as a day with K p ≤ 3 from 00:00 UT to 24:00 UT if the previous day was a day with K p ≤ 3 from 00:00 UT to 24:00 UT.Furthermore, we only use the quiet day with the uninterrupted fof2 measurements from 00:00 UT to 24:00 UT, and the comparison between fof2(d) and fof2(q) measured at the chosen station is carried out if the time difference between fof2(d) and fof2(q) measurements is less than or equal to 30 days.We use the nearest quiet day to the studied disturbed time period, and determine the relative deviation, δ, of fof2 observed at the given station from fof2(q) as Negative and positive values of δ correspond to negative and positive disturbances in NmF2, respectively.We study the dependence of the probabilities of the negative disturbance occurrences in NmF2 on χ.Following the preceding work by Lobzin and Pavlov (2002), we give negative fof2 disturbances the labels "weak" (−0.1 < δ < 0 or 0.81<NmF2(d)/NmF2(q)<1), "normal" (−0.3 < δ ≤ −0.1 or 0.49<NmF2(d)/NmF2(q)≤ 0.81), "strong" (0.5 < δ ≤ −0.3 or 0.25<NmF2(d)/NmF2(q)≤0.49),and "very strong" (δ ≤ −0.5 or NmF2(d)/NmF2(q)≤ 0.25), and confine our attention to relationships between them and the G condition or F1-layer occurrences.
Similar to the G (χ ) and F 1 (χ) determinations, we analyze the measured fof1 and fof2 within each abovedetermined latitude, month and local time range.We determine the probability, δ1≤δ≤δ0 (χ), of the NmF2 negative disturbance occurrence as a ratio of a number of NmF2 negative disturbance observations within the δ1 ≤ δ ≤ δ0 range for zenith angles within the χ interval to a total number of studied NmF2 negative and positive disturbance observations for the same χ, within the given latitude, month, and local time range.

F1-layer and G condition occurrence probabilities
Similar to the preceding work by Lobzin and Pavlov (2002), the total number of hourly measurements studied is 20 532 879 which includes 69 443 G condition occurrences and 2 711 074 F1-layer occurrences.Our NmF2 disturbance analysis includes only negative and positive ionospheric disturbances that have reference quiet days (see Sect. 3).A part of the hourly fof2 disturbance measurements has no reference quiet days, in agreement with the quiet day definition accepted in our paper, and these hourly fof2 measurements are not analyzed.
Average values of F 1 (χ) and G (χ ) are presented in Table 1.For each studied latitude range and season, the first number is determined as an average value, < F 1 (χ ) > 1 , of F 1 (χ) or an average value, < G (χ ) > 1 , of G (χ ) for the first half of a day for χ ≤ 90 • , while the second number is determined as an average value, < F 1 (χ ) > 2 , of F 1 (χ ) or an average value, < G (χ ) > 2 , of G (χ ) for the second half of a day for χ ≤ 90 • .Table 1 shows that < F 1 (χ ) > 2 is less than < F 1 (χ ) > 1 , except for latitude range 5 during the winter, spring, and autumn months.Our calculations show that < G (χ) > 2 is less than < G (χ ) > 1 in latitude ranges 2-5 for all seasons, except for latitude range 3 in winter.For latitude range 3, < G (χ ) > 2 is approximately equal to < G (χ) > 1 for all seasons.
The dependencies of the F1-layer percentage occurrences on the solar zenith angle in latitude range 1 are shown in Fig. 1 for the local time period from 00:00 LT to 12:00 LT (left panel) and from 12:00 LT to 24:00 LT (right panel).Figures 2-5 show the dependence of the F1-layer (bottom panels) and G condition (top panels) percentage occurrence on the solar zenith angle in latitude range 2 (Fig. 2), in latitude range 3 (Fig. 3), in latitude range 4 (Fig. 4), and in latitude range 5 (Fig. 5), during the winter (solid lines), summer (dashed lines), and spring and autumn (dotted lines) months.The left panels of Figs.2-5 represent the F1-layer and G condition percentage occurrence from 00:00 LT to 12:00 LT, while the right panels of Figs.2-5 give the F1-layer and G condition percentage occurrence from 12:00 LT to 24:00 LT.It should be noted that the calculated value of the G condition occurrence probability is negligible above the geomagnetic equatorial region (latitude range 1) during all seasons and in latitude range 2 in winter.Therefore, the G (χ ) dependencies are not discussed in this work for these cases.
Figures 2-5 show that the G condition is more likely to occur during the first half of a day than during the second half of a day in latitude ranges 2-5 during all seasons for the same value of the solar zenith angle, except for latitude range 3 in winter, when the G condition occurrence probability is approximately the same for the same solar zenith angle before and after 12:00 LT.The F1-layer occurrence probability is larger in the first half of a day in comparison with that in the second half of a day for the same value of the solar zenith angle within latitude range 1 for all seasons, while the F1layer occurrence probability is approximately the same for the same solar zenith angle before and after 12:00 LT in latitude ranges 4 and 5.
It can be seen from Figs. 1-5 that the F1-layer and G condition are more commonly formed near midday than near post sunrise or pre-sunset, when the F-region is in the sunlight.These results are in agreement with the conclusions of previous studies (for more details, see, for example, Ratcliffe, 1956Ratcliffe, , 1972;;Yonezawa et. al., 1959;Polyakov et al., 1968) based on the theoretical studies and the limited data set.Figures 1-5 provide for the first time the quantitative measure of the probability variations with solar zenith angle changes.Figures 1-4 show that the maximum values of the F1-layer and G condition occurrence probabilities are located in the 0 • −45 • solar zenith angle range in latitude ranges 1-4.The maximum values of the F1-layer and G condition occurrence probabilities are realized for the minimum value of χ close to noon in latitude range 5 (see Fig. 5).At the minimum solar zenith angle values (see Figs. 2-3), the number of observations is large enough for the identifiable oscillations in the probabilities at these solar zenith angles to be statistically significant.The physical reasons for the occurence of these oscillations are unclear.
The comparison in the values of F 1 (χ ) and G (χ ) between the five latitude regions described above shows the daytime tendency for a decrease in these probabilities at low latitudes and an increase in these probabilities at high latitudes for all seasons.) during the winter (solid lines), summer (dashed lines), and spring and autumn (dotted lines) months before midday from 00:00 LT to 12:00 LT (left panels), and after midday from 12:00 LT to 24:00 LT (right panels).
In the previous F1-layer and G condition studies (Ratcliffe, 1956(Ratcliffe, , 1972;;Polyakov et al., 1968;Shchepkin et al., 1984) based on the limited data set, it was demonstrated that the chance that the F1-layer and G condition will be formed is greater in summer than in winter.Lobzin and Pavlov (2002) have provided additional evidence of this phenomenon by calculating for the first time the diurnally average F1-layer and G condition probability variations with season.Comparison between solid (winter months), dashed (summer months), and dotted (spring and autumn months) lines in Figs.2-5 gives a more detailed picture of the F1layer and G condition seasonal probability behavior for the given solar zenith angle in each latitude ranges 1-5, thereby providing evidence that the chance that the daytime F1-layer and G condition will be formed is greater in summer than in winter.We have found for the first time that the F1-layer occurrence probability is greater in winter than in summer for all solar zenith angles above the geomagnetic equatorial region (see Fig. 1).It should also be noted that the F1layer and G condition seasonal probabilities are lower during the spring and autumn months as compared with that during ) during the winter (solid lines), summer (dashed lines), and spring and autumn (dotted lines) months before midday from 00:00 LT to 12:00 LT (left panels), and after midday from 12:00 LT to 24:00 LT (right panels).
the summer months for most of the solar zenith angle range in latitude ranges 3-5 (compare dotted and dashed lines of Figs.3-5).Scotto et al. (1998) found that the probability of evaluating the occurrence of the F1-layer can be presented as where K = 6.42182 − 0.00252479 2 + 4.02531 × 10 −7 4 , and the unit of P S is percent.The analysis of Scotto et al. (1998) was based on data acquired by the Ionospheric Digital Database of the National Geophysical Data Center, Boulder, Colorado, from 1969 to 1990.The value of P S ( , χ ) approaches 100% for χ = 0 • and 0% for χ = 90 • .
By comparing the dependence of the F1-layer occurrence probability on the solar zenith angle found in our work in each latitude range with P S ( , χ ), we conclude that P S ( , χ ) overestimates the real value of the F1-layer occurrence probability shown in Figs.1-5.For example, 80% ranges 1-3 with the F1-layer occurrence probability shown in Figs.1-3.It is also unclear why there are no seasonal differences in the dependence of P S on χ .The details of deriving Eq. ( 4) are not presented in the short paper by Scotto et al. (1998), and we cannot give an explanation of the identifiable differences between our results and P S ( , χ).

NmF2 normal, strong and very strong negative disturbance occurrence probabilities
Average values of all studied NmF2 negative disturbance probabilities calculated in all latitude ranges and during all seasons are presented in Table 2.For each sort of NmF2 negative disturbance, studied latitude range and season, the first number is determined as an average value, < δ1≤δ≤δ0 (χ) > 1 of δ1≤δ≤δ0 (χ) for the first half of a day for χ ≤ 90 • , while the second number is determined as an average value, < δ1≤δ≤δ0 (χ ) > 2 , of δ1≤δ≤δ0 (χ) for the second half of a day for χ ≤ 90 • .The third number is determined as an average value of δ1≤δ≤δ0 (χ) for the night-time period for χ > 90 • .An average daytime value of any NmF2 negative disturbance probability is calculated as a half-sum of the first and second numbers given in Table 2.
Table 2 and Figs.6 and 7 show that the NmF2 normal, strong, and very strong negative disturbances are more frequent on average at night than by day in latitude ranges 1 and 2 for all seasons, reaching their maximum and minimum occurrence probability values at night and by day, respectively.This conclusion is also correct for all other studied latitude regions during the winter months (see Table 2 and solid lines in Figs.8-10), except for the NmF2 normal and strong negative disturbances in latitude range 5 (see Figs. 10b, c).Table 2 and the dashed and dotted lines in Figs.9c, d and Figs. 10c,  d show that the average night-time occurrence probability is larger than the average daytime occurrence probability for the strong and very strong negative disturbances in latitude ranges 4 and 5 during the summer, spring, and autumn months.
The Joule heating of the thermosphere can be viewed as the frictional heating produced in the thermosphere as the rapidly convecting ions collide with neutral molecules.Most of the Joule heating is deposited in the 115-150 km altitude region, although some extends to higher altitudes (Richmond and Lu, 2000).The geomagnetic storm Joule heating of the thermosphere is considerably more effective than the energy of the auroral electrons in affecting the thermospheric circulation and in the increase in the neutral temperature (Richmond and Lu, 2000).Joule heating from the dissipation of ionospheric currents raises the neutral temperature of the upper thermosphere, and the ion drag drives the high-velocity neutral winds during geomagnetic storms at high latitudes (Prölss, 1980(Prölss, , 1995;;Fuller-Rowell et al., 1996, 2000).It leads to generation of a disturbed composition zone of the high-latitude neutral atmosphere, with an increase in heavier gases and a decrease in lighter gases, i.e. with an increase in The wind surge propagates from aurora regions to low latitudes in both hemispheres.As a result, thermospheric altitude distributions of neutral species at middle and low latitudes are influenced by a global, large-scale wind circulation which is produced by a geomagnetic storm energy input at high latitudes (theoretical and observational studies of thermospheric composition responses to the transport of neutral species from auroral regions to middle latitudes during geomagnetic storms are reviewed by Prölss, 1980Prölss, , 1995)).The increase in the [N 2 ]/[O] ratio maximises in a region that is roughly located in the vicinity of the auroral oval, and this [N 2 ]/[O] increase intensifies and can expand to middle magnetic latitudes with the K p increase (Brunelli and Namgaladze, 1988;Prölss, 1980Prölss, , 1995;;Zuzic et al., 1997;Buonsanto, 1999).The high-latitude geomagnetic storm upwelling brings air rich in the heavy species N2 and O2 to high altitudes, and the geomagnetic storm circulation carries this N 2 and O 2 -rich air to mid-latitudes and lower latitudes.The geomagnetic storm downwelling leads to the opposite effect: air with low values of [N 2 ] and [O 2 ] is carried downward, reducing their concentrations at all altitudes (e.g.Fuller-Rowell et al., 1996;Field et al., 1998;Richmond and Lu, 2000).Thus, the values of transport of neutral composition changes to lower latitudes.The resulting equatorward wind is stronger at night, because the additional geomagnetic storm equatorward wind is added to the quiet day-to-night circulation and because the additional wind is reinforced by antisunward ion drag due to magnetospheric convection E×B drifts (Straus and Schulz, 1976;Babcock and Evans, 1979).As a result, the neutral composition disturbance zone reaches more lower latitudes at night than by day, and the NmF2 normal, strong and very strong negative disturbances tend to be more frequent on average at night than by day in latitude ranges 1 and 2 for all seasons (see Figs. 6 and 7).On the other hand, a rise in hmF2 to regions with a reduced loss rate of O + ( 4 S) ions due to equatorward winds produces an increase in NmF2, while a drop in hmF2 due to poleward winds reduces NmF2.This competition between a neutral composition disturbance causing NmF2 negative storm effects and a rise in hmF2 causing NmF2 positive storm effects determines the complicated dependence of the normal, strong and very strong NmF2 neg- ative disturbance percentage occurrences on the solar zenith angle (see Figs. 6-10).
We found that there is a difference between the dependence of the strong and very strong NmF2 negative disturbance percentage occurrences on the solar zenith angle in latitude ranges 1 and 2 (see Figs. 6c,d and Figs. 7c,d).There are clear discriminated peaks in the probabilities of the strong and very strong NmF2 disturbances before sunrise for all seasons in latitude range 1, while the strong and very strong NmF2 negative disturbance occurrence probabilities are decreased (with some oscillations) with the decrease in the solar zenith angle in the post midnight night-time sector in latitude range 2. Our calculations show (see Figs. 9c,d and Figs. 10c,d) that the −0.5<δ≤−0.3(χ ) peak and the δ≤−0.5 (χ ) peak in the post midnight night-time sector are accompanied by the −0.5<δ≤−0.3(χ ) peak and the δ≤−0.5 (χ ) peak in the sunset-to-midnight sector in latitude ranges 4 and 5 during the winter, spring, and autumn months.

NmF2 weak negative disturbance occurrence probabilities
Average values of the weak NmF2 negative disturbance probabilities calculated for all latitude ranges and seasons are presented in Table 2.For each studied latitude range and season, the first number is determined as an average value, < −0.1<δ<0 (χ) > 1 , of < −0.1<δ<0 (χ ) for the first half of a day for χ ≤ 90 • , while the second number is determined as an average value, < −0.1<δ<0 (χ ) > 2 , of < −0.1<δ<0 (χ ) for the second half of a day for χ ≤ 90 • .The third number is determined as an average value of −0.1<δ<0 (χ ) for the night-time period for χ > 90 • .An average daytime value of the weak NmF2 negative disturbance probability is calculated as a half-sum of the first and second numbers given in Table 2, shows that < −0.1<δ<0 (χ ) > 2 is less than < −0.1<δ<0 (χ ) > 1 in latitude ranges 3-5 for all seasons, except for latitude range 3 in winter and latitude range 5 for the spring and autumn months.In opposition to latitude ranges 3-5, where < −0.1<δ<0 (χ ) > 1 is less than < −0.1<δ<0 (χ ) > 2 in latitude ranges 1 and 2 for all seasons, except for latitude range 2 for spring and autumn months.
The top panels of Figs.6-10 show that the occurrence probability of the NmF2 weak negative disturbances reaches its maximum and minimum values during daytime and nighttime conditions, respectively.It also follows from Table 2 that the average night-time value of this probability is less than that by day for all seasons in all latitude regions.
It should be noted that the F2-layer reaction to weak storms and to substorms is not easily directly observable, since weak effects are masked by ionization variations or by other variations that are not related to geomagnetic activity (hour-to-hour and day-to-day variability, etc.).The variations in the neutral atmosphere, the neutral winds, and the solar EUV flux can be reflected in the variability of NmF2, and it is ascertained that night-time variability is larger than daytime variability of NmF2 (Forbes et al., 2000).Published values of ionospheric electron content were used by Aravindakshan and Iyer (1993) to study its day-to-day variability at a number of stations extending from equatorial to midlatitudes in Indian and American sectors for high and low solar activity years.The variability is larger at night than by day, highest in February and November and lowest in equinox months (Aravindakshan and Iyer, 1993).As a result, we conclude that the identifiable greater probability of the NmF2 weak negative disturbances by day than at night (see the top panels of Figs.6-10) is not related with a variability in the ionosphere.
In addition to the modified large-scale circulation of the neutral atmosphere, during geomagnetic disturbances, the spatial and temporal variations of high-latitude thermosphere heat sources excite large amplitude gravity waves, which produce travelling ionospheric disturbances in the F-region of the ionosphere (Millward et al., 1993;Hocke and Schlegel, 1996).Such gravity waves propagate from high to low latitudes considerably faster in the thermosphere than typical mid-and low-latitude winds resulting from storms (Rees, 1995;Hocke and Schlegel, 1996).The response of the midlatitude ionosphere to this gravity waves propagation is observed by ionozonde stations and incoherent scatter radars in the raising or lowering of hmF2, often by several 10s of km, leading to a decrease or an increase in L and [O] at hmF2, i.e. leading to the increase or decrease in NmF2, respectively (Rees, 1995;Hocke and Schlegel, 1996).The analysis of fof2 measurements shows that night-time fof2 decreases due to gravity wave propagation are not so significant as by day (Deminova et al., 1998).As a result, we conclude that the identifiable greater probability of the NmF2 weak negative disturbances by day than at night, shown in the top panels of Figs.6-10, can be explained if we suggest that NmF2 weak negative disturbances are created by gravity wave propagation in the ionosphere.
4.4 Relationships between the F1-layer, NmF2 negative disturbance, and G condition occurrence probability dependencies on χ Figures 2-5 show that the daytime dependence of the F1layer occurrence probability on the solar zenith angle is generally in phase with that for the G condition in latitude ranges 3-5 for all seasons, and in latitude range 2 during the spring, summer, and autumn months.However, we can conclude from Fig. 2 that this daytime coupling is less convincing in latitude range 2 during the winter months.The competition between the F1 and F2-layers for density dominance determines the G condition occurrence probability.Therefore, the occurrence probabilities of the weak, normal, strong, and very strong NmF2 negative disturbances, in addition to the F1-layer occurrence probability, must be considered in addressing the causes of G condition solar zenith angle changes and in studying the possible relationships between the F1layer and NmF2 negative disturbance occurrences.The G condition in the geomagnetically disturbed ionosphere is associated mainly with a significant negative ionospheric storm in NmF2 (Lobzin and Pavlov, 2002).Thus, the found F1layer occurrence probability dependence on the solar zenith angle (see the low panels of Figs.2-5) and the identifiable solar zenith angle trends in strong and very strong negative disturbance probabilities shown in panels A decrease in the solar zenith angle leads to decreases in the daytime values of strong and very strong negative disturbance occurrence probabilities in latitude ranges 4 and 5 for the spring, summer, and autumn months (dashed and dotted lines in panels (c) and (d) of Figs. 9 and 10).The daytime probabilities −0.3≤δ<−0.1 (χ ), and δ≤−0.5 (χ ) of the NmF2 negative disturbance occurrence do not show discriminated trends in an increase or a decrease with the solar zenith angle decrease in latitude range 2 for all seasons (panels (c) and (d) of Fig. 7), and in all the studied latitude regions for the winter months (solid lines in panels (c) and (d) of Figs.6-10).This means that that the dependence of the F1-layer occurrence probability on χ is the main source which contributes to the daytime G (χ ) trend in the all studied latitude regions for the winter months, in latitude range 2 for all the seasons, and in latitude ranges 4 and 5 for spring, summer, and autumn months.
The results presented in panels (c) and (d) of Fig. 7 show that the occurrence probabilities of the strong and very strong negative disturbances in latitude range 3 is increased with some oscillations if the solar zenith angle is increased.On the other hand, if we do not take into consideration the 0 • − 15 • solar zenith angle range during the summer, spring, and autumn months, then we can conclude that the F1-layer occurrence probability is decreased in latitude range 3 if the solar zenith angle is increased (see Fig. 3).This means that the solar zenith angle trend in the G condition occurrence probability arises in the main from the solar zenith angle trend in the F1-layer occurrence probability of this latitude region.The solar zenith angle trend in the probabilities of strong and very strong NmF2 negative disturbances counteracts the solar zenith angle trend in the probability of the G condition occurrence shown in the upper panels b and c of Fig. 3.

Conclusions
The primary goal of the present work is to calculate the dependencies of the NmF2 negative disturbance, F1-layer and G condition occurrence probabilities on the solar zenith angle during the summer, winter, spring and autumn months  Center, Boulder, Colorado, from 1957to 1990.The G condition cannot exist in the ionosphere if there is no F1-layer.During ionospheric disturbances, the NmF2 decrease leads to the increase in the G condition occurrence probability if the F1-layer exists.The relationships between the G condition, F1-layer, and NmF2 negative disturbance occurrence probabilities are also studied in this paper.

F1-Layer and the G condition
Our calculations show that the G condition is more likely to occur during the first half of a day than during the second half of a day, in latitude ranges 2-5 during all seasons for the same value of the solar zenith angle, except for latitude range 3 in winter, when the G condition occurrence probability is approximately the same for the same solar zenith angle before and after 12:00 LT for χ ≤ 90 • .
We found that the average value for the second half of a day of the F1-layer occurrence probability is less than that for the first half of a day, except for latitude range 5 during the winter, spring, and autumn months.The F1-layer occurrence probability is larger in the first half of a day in comparison with that in the second half of a day for the same value of the solar zenith angle in latitude range 1 for all seasons, while the F1-layer occurrence probability is approximately the same for the same solar zenith angle before and after noon in latitude ranges 4 and 5 for χ ≤ 90 • .
We found that the F1-layer and G condition are more commonly formed near midday than close to post sunrise or presunset, when the F-region is in the sunlight.The maximum values of the F1-layer and G condition occurrence probabilities are found to be in the 0 • − 45 • solar zenith angle range in latitude ranges 1-4.The maximum values of the F1-layer and G condition occurrence probabilities are realized for the minimum value of χ close to noon in latitude range 5.
The comparison in the values of the F1-layer and G condition occurrence probabilities between all studied latitude ranges shows the daytime tendency for a decrease in these probabilities at low geomagnetic latitudes and an increase in these probabilities at high geomagnetic latitudes for all seasons.The identifiable detailed picture of the F1-layer and G condition seasonal probability behavior at the given solar zenith angle in latitude ranges 2-5 provides evidence that the chance that the daytime F1-layer and G condition will be formed is greater in summer than in winter.We have found for the first time that the F1-layer occurrence probability is greater in winter than in summer for all solar zenith angles in latitude range 1.The identifiable F1-layer and G condition seasonal probabilities are lower during the spring and autumn months as compared with that during the summer months for most of the solar zenith angle range in latitude ranges 3-5.

Negative NmF2 disturbances
The magnitudes of the studied NmF2 weak, normal, strong and very strong negative disturbances and their extension to lower latitudes are controlled by a number of parameters, including the strength of the magnetospheric storm or substorm, the season, the latitude, and the solar zenith angle.Our results clearly capture the geomagnetic latitude dependence in the NmF2 normal, strong, and very strong negative disturbance probabilities, reproducing the general tendency for a decrease in these probabilities at low latitudes and an increase in the probabilities at high latitudes.We found that the NmF2 normal, strong, and very strong negative disturbances are more frequent on average at night than by day, in latitude ranges 1 and 2 for all seasons, reaching their maximum and minimum occurrence probability values at night and by day, respectively.This conclusion is also correct for all other studied latitude regions during the winter months, except for the NmF2 normal and strong negative disturbances in latitude range 5.The calculated average night-time occurrence probability is larger than the average daytime occurrence probability for the strong and very strong negative disturbances in latitude ranges 4 and 5 during the summer, spring, and autumn months.
It is proved that the average value for the second half of a day of the NmF2 negative disturbance probability is less than that for the first half of a day for the normal, strong, and very strong NmF2 negative disturbances in latitude ranges 3-5 for all seasons, except for the very strong NmF2 negative disturbances in latitude ranges 3 in winter and in latitude range 5 during the winter, spring, and autumn months, when the average value of the very strong NmF2 negative disturbances for χ ≤ 90 • is approximately the same before and after 12:00 LT.In opposition to latitude ranges 3-5, the average value for the first half of a day of the NmF2 negative disturbance probability is less than that for the second half of a day for the normal, strong, and very strong NmF2 negative disturbances in latitude ranges 1 and 2 in winter, except for the very strong NmF2 negative disturbances in latitude range 2 during all seasons and that in latitude range 1 during the summer and winter months.The average value of the very strong NmF2 negative disturbances for χ ≤ 90 • is approximately the same before and after 12:00 LT in latitude range 2.
We found that there is a difference between the dependence of the strong and very strong NmF2 negative disturbance percentage occurrences on the solar zenith angle in latitude ranges 1 and 2. There is a clear, discriminated peak in the probability of the strong or very strong NmF2 disturbance before sunrise for all seasons in latitude range 1.The strong and very strong NmF2 negative disturbance occurrence probabilities are decreased (with some oscillations) with the decrease in the solar zenith angle in the post midnight sector in latitude range 2. Our calculations show that the strong and very strong NmF2 negative disturbance occurrence probability peaks in the post midnight night-time sector are accompanied by the peaks in these probabilities in the sunset-tomidnight sector in latitude ranges 4 and 5 during the winter, spring, and autumn months.
It is proved that that the average value for the second half of a day of the weak NmF2 negative disturbance probability is less than that for the first half of a day in latitude ranges 3-5 for all seasons, except for latitude range 3 in winter and latitude range 5 for the spring and autumn months.In opposition to latitude ranges 3-5, the average value for the first half of a day of the weak NmF2 negative disturbance probability is less than that for the second half of a day in latitude ranges 1 and 2 for all seasons, except for latitude range 2 for the spring and autumn months.
The calculated occurrence probability of the NmF2 weak negative disturbances reaches its maximum and minimum values during daytime and night-time conditions, respectively, and the average night-time value of this probability is less than that by day for all seasons in all latitude regions.It is proved that an ionosphere variability that is not related to geomagnetic activity is not the cause of the NmF2 weak negative disturbance probability variations.We have concluded that the identifiable greater probability of the NmF2 weak negative disturbances by day than at night can be explained if we suggest that NmF2 weak negative disturbances are created by gravity wave propagation in the ionosphere.

Relationships between the G condition, F1-layer, and
NmF2 negative disturbance occurrence probabilities The competition between the F1-and F2-layers for density dominance determines the G condition occurrence probability.We found that the daytime dependence of the F1-layer occurrence probability on the solar zenith angle is generally in phase with that for the G condition in latitude ranges 3-5 for all seasons, and in latitude range 2 during the spring, summer, and autumn months, but this daytime coupling is less convincing in latitude range 1 during the winter months.
On the other hand, the G condition in the geomagnetically disturbed ionosphere is associated mainly with a significant negative ionospheric storm in NmF2.Thus, the identifiable F1-layer occurrence probability dependence on the solar zenith angle and the identifiable solar zenith angle trends in strong and very strong negative disturbance probabilities are the trends involved in the formation of the G condition solar zenith angle trends.Our calculations show that the main source which contributes to the daytime dependence of the G condition occurrence probability on the solar zenith angle in the all studied latitude regions for the winter months, in latitude range 2 for all seasons, and in latitude ranges 4 and 5 for the spring, summer, and autumn months, is the dependence of the F1-layer occurrence probability on the solar zenith angle.The solar zenith angle trend in the probability of the G condition occurrence in latitude range 3 arises in the main from the solar zenith angle trend in the F1-layer occurrence probability.The solar zenith angle trend in the probabilities of strong and very strong NmF2 negative disturbances counteracts the identifiable solar zenith angle trend in the probability of the G condition occurrence.

Fig. 10 .
Fig. 10 Figures 2-5show that the daytime dependence of the F1layer occurrence probability on the solar zenith angle is generally in phase with that for the G condition in latitude ranges 3-5 for all seasons, and in latitude range 2 during the spring, summer, and autumn months.However, we can conclude from Fig.2that this daytime coupling is less convincing in latitude range 2 during the winter months.The competition between the F1 and F2-layers for density dominance determines the G condition occurrence probability.Therefore, the occurrence probabilities of the weak, normal, strong, and very strong NmF2 negative disturbances, in addition to the F1-layer occurrence probability, must be considered in addressing the causes of G condition solar zenith angle changes and in studying the possible relationships between the F1layer and NmF2 negative disturbance occurrences.The G condition in the geomagnetically disturbed ionosphere is associated mainly with a significant negative ionospheric storm in NmF2(Lobzin and Pavlov, 2002).Thus, the found F1layer occurrence probability dependence on the solar zenith angle (see the low panels of Figs.2-5) and the identifiable solar zenith angle trends in strong and very strong negative disturbance probabilities shown in panels (c) and (d) of Figs.7-10, are the trends involved in the formation of the G condition solar zenith angle tendencies shown in the top panels of Figs.2-5.A decrease in the solar zenith angle leads to decreases in the daytime values of strong and very strong negative disturbance occurrence probabilities in latitude ranges 4 and 5 for the spring, summer, and autumn months (dashed and dotted lines in panels (c) and (d) of Figs. 9 and 10).The daytime probabilities −0.3≤δ<−0.1 (χ ), and δ≤−0.5 (χ ) of the NmF2 negative disturbance occurrence do not show discriminated trends in an increase or a decrease with the solar zenith angle decrease in latitude range 2 for all seasons (panels (c) and (d) of Fig.7), and in all the studied latitude regions for the winter months (solid lines in panels (c) and (d) of Figs.6-10).This in latitude range 1 (| | ≤ 10 • ), in latitude range 2 (10 • < | | ≤ 30 • ), in latitude range 3 (both 30 • < |ϕ| ≤ 45 • and 30 • < | | ≤ 45 • ), in latitude range 4 ( both 45 • < |ϕ| ≤ 60 • and 45 • < | | ≤ 60 • ), and in latitude range 5 (60 • < | | ≤ 90 • ), using experimental data acquired by the Ionospheric Digital Database of the National Geophysical Data

Table 1 .
Average values of F1-layer and G condition percentage occurrences in latitude ranges 1-5 during the winter, summer, and spring and autumn months for the first (first number) and second (second number) half of the day for χ ≤ 90 • Latitude range 1 Latitude range 2 Latitude range 3 Latitude range 4 Latitude range 5

Table 2 .
Average values of weak, normal, strong, and very strong NmF2 negative disturbance percentage occurrence in latitude ranges 1-5 during the winter, summer, and spring and autumn months for the first (first number) and second (second number) half of a day for χ ≤ 90 • , and during the night-time period for χ > 90 • (third number)