Solar zenith angle dependencies of F1-layer, NmF2 negative disturbance, and G-condition occurrence probabilities

Transcription

1 Annles Geophysie (22) 2: Europen Geosienes Union 22 Annles Geophysie Solr zenith ngle epenenies of F1-lyer, NmF2 negtive isturne, n G-onition ourrene proilities V. V. Lozin n A. V. Pvlov Institute of Terrestril Mgnetism, Ionosphere, n Rio-Wve Propgtion, Russin Aemy of Sienes (IZMIRAN), Troitsk, Mosow Region, 14219, Russi Reeive: 1 Jnury 22 Revise: 5 April 22 Aepte: 28 My 22 Astrt. Experimentl t quire y the Ionospheri Digitl Dtse of the Ntionl Geophysil Dt Center, Bouler, Coloro, from 1957 to 199, re use to stuy the epenene of the G onition, F1-lyer, n NmF2 negtive isturne ourrene proilities on the solr zenith ngle uring summer, winter, spring, n utumn months in ltitue rnge 1 (etween 1 n +1 of the geomgneti ltitue, ), in ltitue rnge 2 (1 < 3 ), in ltitue rnge 3 (3 < ϕ 45, 3 < 45 ), in ltitue rnge 4 (45 < ϕ 6, 45 < 6 ), n in ltitue rnge 5 (6 < 9 ), where ϕ is the geogrphi ltitue. Our lultions show tht the G onition is more likely to our uring the first hlf of y thn uring the seon hlf of y, t ll ltitues uring ll sesons for the sme vlue of the solr zenith ngle. The F1-lyer ourrene proility is lrger in the first hlf of y in omprison with tht in the seon hlf of y for the sme vlue of the solr zenith ngle in ltitue rnge 1 for ll sesons, while the F1-lyer ourrene proility is pproximtely the sme for the sme solr zenith ngle efore n fter noon in ltitue rnges 4 n 5. The F1-lyer n G onition re more ommonly forme ner miy thn lose to post sunrise or pre-sunset. The hne tht the ytime F1-lyer n G onition will e forme is greter in summer thn in winter t the given solr zenith ngle in ltitue rnges 2 5, while the F1-lyer ourrene proility is greter in winter thn in summer for ny solr zenith ngle in ltitue rnge 1. The lulte ourrene proility of the NmF2 wek negtive isturnes rehes its mximum n minimum vlues uring ytime n night-time onitions, respetively, n the verge night-time vlue of this proility is less thn tht y y for ll sesons in ll stuie ltitue regions. It is shown tht the NmF2 norml, strong, n very strong negtive isturnes re more frequent on verge t night thn y y in ltitue rnges 1 n 2 for ll sesons, rehing their mximum n minimum ourrene proility vlues t night n y y, respe- Corresponene to: A. V. Pvlov (pvlov@izmirn.rssi.ru) tively. This onlusion is lso orret for ll other stuie ltitue regions uring winter months, exept for the NmF2 norml n strong negtive isturnes in ltitue rnge 5. A ifferene in the epenene of the strong n very strong NmF2 negtive isturne perentge ourrenes on the solr zenith ngle is foun etween ltitue rnges 1 n 2. Our results provie eviene tht the ytime epenene of the G onition ourrene proility on the solr zenith ngle is etermine minly y the epenene of the F1-lyer ourrene proility on the solr zenith ngle in the stuie ltitue regions for winter months, in ltitue rnge 2 for ll sesons, n in ltitue rnges 4 n 5 for spring, summer, n utumn months. The solr zenith ngle tren in the proility of the G onition ourrene in ltitue rnge 3 rises in the min from the solr zenith ngle tren in the F1-lyer ourrene proility. The solr zenith ngle tren in the proilities of strong n very strong NmF2 negtive isturnes ounterts the ientifie solr zenith ngle tren in the proility of the G onition ourrene. Key wors. Ionosphere (ionospheri isturnes, ionosphere-tmosphere intertions, ion hemistry n omposition) 1 Introution The Ionospheri Digitl Dtse of the Ntionl Geophysil Dt Center, Bouler, Coloro, provies the routine souning groun-se sttion mesurements of the ritil frequenies n virtul heights of ifferent ionospheri lyers, n, in prtiulr, the ritil frequenies fof1 n fof2 of F1- n F2-lyers tht re nlyze in this stuy. The vlues of the pek ensities, NmF1 n NmF2, of the F1- n F2-lyers re relte to the ritil frequenies fof2 n fof1 s NmF2= fof2 2 n NmF1= fof1 2, where the unit of NmF2 n NmF1 is m 3, the unit of fof2 n fof1 is MHz (URSI hnook of ionogrm interprettion n reution, 1978). The Ionospheri Digitl Dtse is

2 1822 V. V. Lozin n A. V. Pvlov: Solr zenith ngle epenenies forme using the URSI stnr rules (URSI hnook of ionogrm interprettion n reution, 1978). In ition to numeril vlues of ionospheri prmeters, the qulifying n esriptive letters A Z re use in this tse. The esriptive letter G mens tht mesurement is influene y, or impossile euse, the ioniztion ensity of the lyer is too smll to enle it to e me urtely, n this se is esrie s G onition in the F-region of the ionosphere when fof2 fof1 (URSI hnook of ionogrm interprettion n reution, 1978). If the lyer is not seen from ionogrms ue to other resons, then other letters re use. The G onition rises in the ionosphere when the ritil frequeny of the F2-lyer rops elow tht of the F1-lyer, i.e. when the pek ensity, NmF1, of the F1-lyer, whih is ompose mostly of the moleulr ions NO + n O2 +, is lrger thn tht of the F2-lyer, whih is ominte y O + ions (King, 1962). As result, very low min pek ltitue vlue (elow 2 km) is oserve in ionogrms, so tht no informtion is otinle ove this height from groun-se ionozone t. As fr s the uthors know, the first ltitue istriution of the eletron ensity uring G onition ws eue y Norton (1969) from ionogrms reore y the Alouette I stellite ionozone n the St. John s groun-se ionozone uring the severe negtive ionospheri storm on 18 April The physis of the G onition phenomenon hs een stuie y Buonsnto (199) using ionosone t from two mi-ltitue sttions, y Oliver (199), using Millstone Hill inoherent stter rr t, n y Fuko et l. (1991) using t from the mile n upper tmosphere rr in Jpn. Pvlov n Buonsnto (1998), Pvlov (1998), Pvlov et l. (1999), Shlesier n Buonsnto (1999), n Pvlov n Foster (21) stuie the G onition formtion for quiet n isture mi-ltitue ionosphere uring perios of low, moerte, n high solr tivity, using the Millstone Hill inoherent stter rr t. Moel results lso show tht O + n eome minor ion in the F-region, reting G onition uring isture onitions t high ltitues (Bnks et l., 1974; Shunk et l., 1975), n oservtions t EIS- CAT onfirm this onlusion (e.g. Häggström n Collis, 199). These ppers provie eviene tht hnges in [O], [N 2 ], [O 2 ], n the plsm rift veloity, the effet of the perpeniulr (with respet to the geomgneti fiel) omponent of the eletri fiel on the eletron ensity (through hnges in the rte oeffiients of hemil retions of ions), n the effets of virtionlly exite N 2 n O 2 on the eletron ensity re importnt ftors tht ontrol the G onition formtion in the ionosphere. The stuy of the G onition formtion in the ionosphere ove Millstone Hill uring the severe geomgneti storm of July 2, provie weighty rgument for the inlusion of the effets of virtionlly exite N 2 n O 2 on the eletron ensity n temperture in ionospheri moels (Pvlov n Foster, 21). During NmF2 isturnes, whih re elieve to e use y geomgneti storms n sustorms, the vlue of NmF2 n either inrese or erese in omprison with geomgnetilly quiet NmF2, n these hnges re enote s positive n negtive isturnes (Prölss, 1995; Buonsnto, 1999). A erese in NmF2 uring NmF2 negtive isturne les to n inrese in the G onition ourrene proility if the F1-lyer exists. On the other hn, the G onition nnot exist in the ionosphere if there is no F1-lyer. The preeing work y Lozin n Pvlov (22) summrizes ppers ressing the mesurements n the physis of the F1-lyer, NmF2 negtive isturne, n G onition, n gives for the first time the etile epenenies of the proilities of the F1-lyer, NmF2 negtive isturne, n G onition ourrenes on ily solr tivity inex, F1.7, 3-h geomgneti inex, K p, numer of given y in yer, n geomgneti ltitue. The im of this pper is to rry out sttistil stuy of solr zenith ngle epenenies of NmF1, NmF2 negtive isturne, n G onition ourrene proilities using the Digitl Dtse fof1 n fof2 mesurements from 1957 to 199. Some fetures of the solr zenith ngle epenenies of NmF1, NmF2 negtive isturne, n G onition ourrene frequenies hve een known for long time. Rtliffe (1956, 1972), Yonezw et l. (1959) onlue tht the F1- lyer is less lile to pper for lrger vlues of the solr zenith ngle. Du Chrme n Petrie (1973) erive n expression to preit fof1, ssuming limits for the presene of the F1-lyer s funtion of the solr zenith ngle n of solr tivity. Sotto et l. (1997) teste the Du Chrme n Petrie (1973) formul opte in the Interntionl Referene Ionosphere (IRI) moel, tking into ount lterntive solutions for the prtiulr restritions impose y the IRI for high vlues of solr zenith ngle. New proility funtions to preit the ourrene of the F1-lyer hve een propose y Sotto et l. (1997, 1998) to reple the Du Chrme n Petrie (1973) formul. A negtive F2 ionospheri storm onset t mile ltitues is most frequently oserve in the morning time setor n very rrely in the noon, fternoon, n night-time setors (Prölss, 1995; Buonsnto, 1999). Wrenn et l. (1987) isriminte geomgneti tivity levels s very quiet, quiet, norml, isture, n very isture onitions. The negtive ionospheri storm effet in NmF2 uring norml, isture, n very isture onitions is entere t night for very isture onitions n uring morning hours for norml n isture onitions if the ionozone fof2 mesurements from the Argentine Islns ionozone sttion re use (Wrenn et l., 1987). The omprison etween the summer fof2 mesurements of the Argentine Islns n Port Stnley ionozone sttions les to the onlusion tht the mximum fof2 epression moves from the night-time setor to the morning setor if the ltitue of the sttion is hnge from mile to more low ltitues (Wrenn et l., 1987). It is foun y Rtliffe (1972) tht the G onition is more ommonly forme ner miy thn uring severl hours fter sunrise or efore sunset. However, the results of Rtliffe (1972) re not formulte in mthemtil form to e use in lultions. Previous investigtions re se on the limite ionozone

3 V. V. Lozin n A. V. Pvlov: Solr zenith ngle epenenies 1823 t set from some sttions n on theoretil nlysis of the min physil proesses tht form eletron ensity ltitue profiles. As result, exept for the Sotto et l. (1997, 1998) formuls for the F1-lyer proility funtion, there re no pulishe solr zenith ngle epenenies of NmF1, NmF2 negtive isturne, n G onition ourrene proilities. The min purpose of this work is to lulte for the first time these proilities for low, mile, n high ltitues in summer, in winter, n uring the spring n utumn months, to provie some quntittive mesure of these proility vritions. In our nlysis we stuy for the first time possile reltionship etween the solr zenith ngle proility epenene of the G onition ourrene with the solr zenith ngle proility epenenies of NmF2 negtive isturne n F1-lyer ourrenes. 2 Formtion of the F1- n F2-lyers in the ionosphere Solr zenith ngle epenenies of NmF1 n NmF2 negtive isturne n G onition ourrene proilities, whih re stuie in our work, re etermine y physil proesses tht form the F-region of the ionosphere. The F-region is lote in the ltitue rnge ove km. Within the F-region re the F1- n F2-lyers, with the pek ltitues hmf1< 19 2 km n hmf2> 2 21 km, respetively. The mjor F1- n F2-lyer ions re O + ( 4 S), O + 2, n NO +. The min physil proesses tht form the F1 n F2-lyers in the ionosphere y lne etween proution, hemil loss, n trnsport of eletrons n ions re esrie in mny review rtiles, ooks, n ppers (e.g. Rtliffe, 1972; Risheth n Grriot, 1969; Brunelli n Nmglze, 1988; Rees, 1989; Fejer, 1997; Risheth n Muller-Worg, 1999; Risheth, 2; Risheth et l., 2; Au, 21; Lozin n Pvlov, 22; Pvlov n Foster, 21). Following these stuies, it is usully suppose tht the vlue of NmF2 is pproximtely iretly proportionl to the [O]/L rtio t hmf2 uring ytime onitions, where L is the loss rte of O + ( 4 S) ions in the retions of O + ( 4 S) with unexite N 2 (v = ) n O 2 (v = ) n virtionlly exite N 2 (v) n O 2 (v) moleules t virtionl levels, v >. Thus, the epletion in [O] n the inrese in [N 2 ] n [O 2 ] n le to negtive phse in NmF2. The inrese in the rte oeffiients for retions etween O + ( 4 S) ions n N 2 (v ) n O 2 (v ), ue to hnges in neutrl n ion tempertures n ue to the inrese in virtionl tempertures of N 2 n O 2 woul lso proue negtive storm effets in NmF2. These ssumptions re use in our stuy in isussions of NmF2 vrition soures n to unerstn resons for solr zenith ngle epenenies of NmF2 negtive isturne n G onition ourrene proilities. To illustrte the si physis involve n to stuy the physil resons for NmF2 negtive isturne ourrene proility nighttime vritions, it is useful to use the nlytil pproximtion of the nighttime mi-ltitue NmF2 given y Krinerg n Tshhilin (1982, 1984) s NmF2(t) NmF2(t ) exp [ (t t )L(t)] +3F (t)t n (t)ν in (t)/ (g [T i (t) + T e (t)]), (1) where t is lol time, NmF2(t ) is the ionospheri eletron ensity for the lol time t orresponing to usk, T n is the exospheri neutrl temperture, T i n T e re ion n eletron tempertures, g is the elertion ue to grvity, ν in is the O + O ollision frequeny, F is the vlue of plsm ion flux flowing from the plsmsphere into the ionosphere, the vlues of L, T i, T e, g, n ν in re hosen t the F2 pek ltitue. One n see from Eq. (1) tht the nighttime F2-region eletron ensity onsists of two prts. The first term esries the role of the ytime ioniztion in the mintenne of the nighttime ionosphere. In this se the F-region woul ey with the hrteristi time L 1 (out severl hours). Sine the loss rte of O + ( 4 S) ions is proportionl to [N 2 ] n [O 2 ], n inrese or erese in [N 2 ] n [O 2 ] t hmf2 ltitues les to erese or n inrese in NmF2, respetively. The noturnl F-region is lso mintine y ownwr flow of ioniztion from the plsmsphere, esrie y the seon term in Eq. (1). In winter, n possily in spring n in utumn, when the night is long enough, the role of the seon term in Eq. (1) inreses efore sunrise, n this term n etermine the mi-ltitue vlue of NmF2. The role of the ion trnsport is less thn the role of the hemil retions of ions with eletrons n neutrl omponents of the upper tmosphere t the F1-lyer ltitues, n the proution n loss rtes of eletrons n ions tht etermine the F1-lyer formtion. To stuy the formtion of the F1-lyer, Rtliffe (1972) ssume tht the min soure of NO + ions is the hemil retion of O + with N 2, n there re only NO + n O + ions. Rtliffe (1972) foun tht the pek of the F1-lyer exists in the ionosphere if the pek ltitue, h, of the totl proution rte of therml eletrons is less thn n ltitue, h t. The vlue of h t is etermine from the onition of K [N 2 ] = α[e], where K is the rte oeffiient for the retion of O + ions with N 2, n α is the rte oeffiient of the issoitive reomintion of NO + ions. Rtliffe (1972) onlue tht the vlue of h t h is erese with the solr tivity level inrese, n the vlue of h t h hs mximum vlue lose to miy. As result, the F1 pek is more lerly in eviene t solr minimum thn t solr mximum, n the F1 pek is more ommonly forme ner miy n in summer (Rtliffe, 1972). Yonezw et l. (1959) rrie out nother simple onsiertion for the F1-lyer to pper s istint lyer. In ition to the equlity of α n the rte oeffiient of the issoitive reomintion of O + 2 ions, the height grient of the tmospheri neutrl omponents, the rte oeffiients of O + ions with N 2 n O 2, n the vlue of α were ssume to e onstnts y Yonezw et l. (1959), to otin the onition of pperne of the F1-lyer s (αq ).5 <.89L(h = h )(os χ) 1.65, (2) where Q is the mximum proution rte of eletron-ion pirs for the Sun y photoioniztion of tomi oxygen t n ltitue, h = h.

4 1824 V. V. Lozin n A. V. Pvlov: Solr zenith ngle epenenies Using Eq. (2), Yonezw et l. (1959) onlue tht the F1-lyer is more lile to pper uring perios of low solr tivity thn uring perios of high solr tivity, n uring the y ner noon thn ner sunrise or sunset. It follows from Eq. (2) tht the NmF1 ourrene proility pprohes 1% if Eq. (2) is vli n % if Eq. (2) is not relize in the ionosphere. The ionosphere t low ltitues is very sensitive to eletri fiels. The ytime low ltitue eletri fiel tht is irete estwr uses the ionosphere to e lifte to high ltitues long mgneti fiel lines, where there is very rpi iffusion of eletrons n ions, n grvity pull the eletrons n ions ownwr n polewr on either sie of the mgneti equtor, so tht low ltitue trough evelops over the equtor, with the F2-lyer ensity mximum to the north n south (Risheth, 2; Au, 21). This F2-lyer struture is usully lle the equtoril or Appleton nomly. If the ytime estwr eletri fiel is strengthene or wekene uring geomgneti storms n sustorms, the F2-lyer ensity mximum move further polewr or equtorwr, n the ionospheri ensity over the equtor is reue or inrese, respetively. It is remrkle tht the upwr rift y y is lne y ownwr rift t night, n hmf2 is lower t night thn y y t low ltitues, while the mile ltitue hmf2 is higher t night thn y y. Aoring to Sterling et l. (1972), the low-ltitue F1-lyer is use y the eletromgneti rift, rther thn y n effet of F1-lyer photohemistry. 3 Dt n metho of t nlysis Ionogrms proue y ionozones re reors tht show vritions of the virtul height of rio wve refletion from the ionosphere s funtion of the rio frequeny, h (f ), within the frequeny n rnge 1 MHz 2 MHz tht is normlly use (URSI hnook of ionogrm interprettion n reution, 1978). The rio wve tht is reflete from the ionosphere level of ioniztion is split into two wves of ifferent polriztion in the Erth s mgneti fiel, therey leing to two sorts of oserve h (f ) urves. These wves re lle the orinry wve (o-moe) n the extrorinry wve (x-moe). There re lso z-moe tres on some ionogrms generte y rio wves whih hve een propgte long the mgneti fiel lines. The moe tres n e ientifie y the frequeny seprtion n y other initions presente in the URSI hnook of ionogrm interprettion n reution (1978). A simple pproh is use to fin pek eletron ensities of the ionosphere from oservtions of h (f ) urves. When the level of the pek eletron ensity in the lyer is rehe, the vlue of h (f ) eomes effetively infinite ( f h ). The frequeny t whih this ours is etermine s the ritil frequeny of the ionospheri lyer. Our nlysis is se on 34 yers of hourly fof2 n fof1 t from 1957 to 199 from sttions ville on the Ionospheri Digitl Dtse of the Ntionl Geophysil Dt Center, Bouler, Coloro. At the hosen ionozone sttion, the solr zenith ngle, χ, is funtion of lol time, geogrphi ltitue, n numer, n, of given y in yer. Therefore, multiple-prmeter sttistis is neee to stuy the solr zenith ngle epenenies of NmF2 negtive isturne, NmF1, n G onition ourrenes. We o not nlyze the ionozone mesurements of fof2 n fof1 in the Northern Hemisphere n the Southern Hemisphere seprtely, ut rry out our sttistil nlysis of solr zenith ngle epenenies of NmF1, NmF2 negtive isturne, n G onition ourrenes seprtely in summer (June, July, n August in the Northern Hemisphere, n Deemer, Jnury, n Ferury in the Southern Hemisphere), in winter (Deemer, Jnury, n Ferury in the Northern Hemisphere, n June, July, n August in the Southern Hemisphere), n uring spring n utumn months (Mrh, April, My, Septemer, Otoer, n Novemer in oth hemispheres). As we hve pointe out, the solr zenith ngle is funtion of geogrphi ltitue, ϕ. Therefore, to stuy the solr zenith ngle epenenies of NmF2 negtive isturne, F1-lyer, n G onition ourrenes, the geogrphi ltitue rnge hs to e tken so tht this rnge is minimize, while the numer of mesurements remins lrge enough to rry out this sttistil stuy. On the other hn, there re signifint ifferenes in physil proesses tht etermine NmF2 negtive isturne, NmF1, n G onition ourrenes t low, mile, n high geomgneti ltitues, (e.g. Rtliffe, 1972; Risheth n Grriot, 1969; Brunelli n Nmglze, 1988; Rees, 1989; Fejer, 1997; Risheth n Muller-Worg, 1999; Risheth et l., 2; Lozin n Pvlov, 22), n these ifferenes n le to ifferenes in the solr zenith ngle epenenies of the stuie events. Therefore, we split the ionozone fof2 n fof1 t set use into five prts. A geomgneti equtoril region ( 1 1 ), where n equtoril ytime NmF2 trough is evelope, is efine in our stuy s ltitue rnge 1. A ltitue rnge 2 is low-ltitue region (1 < 3 ) where ytime NmF2 rests in omprison when equtoril ytime NmF2 re oserve. Ltitue rnges 3 n 4 re mi-ltitue regions. A ltitue rnge 3 inlues oth 3 < ϕ 45 n 3 < 45. A ltitue rnge 4 inlues oth 45 < ϕ 6 n 45 < 6. It is ler from the efinition of these mi-ltitue rnges tht ltitue rnges 3 n 4 re not overlpping. The min ioniztion trough, the ioniztion hole in the polr p roun lol wn, the tongue of ioniztion, n the urorlly proue ioniztion eletron ensity pek in the viinity of the urorl ovl re nturl onsequene of the ifferene n ompetition etween the vrious hemil n trnsport proesses known to e operting in the high-ltitue F-region ionosphere (for more etils, see, e.g. Rees, 1989; Buonsnto, 1999). The ltitue n longitue ounries of these regions show mrke vritions. As result, only verge solr zenith ngle epenenies of NmF1, NmF2 negtive isturne, n G onition proility funtions in ltitue rnge 5 (6 < 9 ) re lulte in this pper. To isriminte etween the

5 V. V. Lozin n A. V. Pvlov: Solr zenith ngle epenenies 1825 Tle 1. Averge vlues of F1-lyer n G onition perentge ourrenes in ltitue rnges 1 5 uring the winter, summer, n spring n utumn months for the first (first numer) n seon (seon numer) hlf of the y for χ 9 Ltitue rnge 1 Ltitue rnge 2 Ltitue rnge 3 Ltitue rnge 4 Ltitue rnge 5 Winter 14.1, , , , , 13.4 F1 Summer 12.2, , , , , 46.7 lyer Spring, 9.1, , , , , 44.1 utumn Winter.1,.1.2,.1.8,.6 G Summer.3,.1.9,.5 1.7,.5 3.8, 2. onition Spring,.5,.3.6,.4 1.4,.6 2.2, 2. utumn morning n evening solr zenith ngle epenenies of the stuie events, we split the rnge of χ 18 into twelve intervls of the sme length, χ, oth from : LT to 12: LT, n from 12: LT to 24: LT t eh ionozone sttion. We onsier the mesure fof1 n fof2 within the oveetermine lol time, ltitue, n month rnge, n etermine the proility, G (χ), or F 1 (χ), of the G onition or F1-lyer ourrene s rtio of the numer of G onition or F1-lyer oservtions for zenith ngles within the χ intervl to totl numer of mesurements for the sme χ n within the given lol time, ltitue, n month rnge. The eletron ensity n either erese or inrese uring geomgnetilly isture onitions, n these eletron ensity hnges re enote s negtive n positive ionospheri isturnes, respetively. To test the effets of geomgneti tivity, we use two ifferent K p lels: isture, for whih we tke K p > 3 n use the pek ensity, NmF2(), n ritil frequeny, fof2(), of the F2-lyer oserve uring the time perios with K p > 3, n quiet, for whih we tke K p 3. The etermintion of the quiet pek ensity, NmF2(q), n ritil frequeny, fof2(q), of the F2-lyer, is ruil for stuies of negtive n positive ionospheri isturnes. Perturtions in the neutrl omposition, temperture, n win t one ltitue re rpily trnsmitte to higher n lower ltitues. However, it tkes time to relx k to n initil stte of the thermosphere, n this thermosphere relxtion etermines the time for the isture ionosphere to relx k to the quiet stte. It mens tht not every fof2 oserve uring the y with K p 3 n e onsiere s fof2(q). The hrteristi time of the neutrl omposition reovery fter storm impulse event rnges from 7 to 12 h on verge (Hein, 1987), while it my nee up to severl ys for ll ltitues own to 12 km in the tmosphere to reover ompletely k to the unisture stte of the tmosphere (Rihmon n Lu, 2). As result of this thermosphere reovery, y with K p 3 from : UT to 24: UT nnot e onsiere quiet y if the previous y ws y with K p > 3 from : UT to 24: UT. We etermine the quiet referene y with fof2(q) s y with K p 3 from : UT to 24: UT if the previous y ws y with K p 3 from : UT to 24: UT. Furthermore, we only use the quiet y with the uninterrupte fof2 mesurements from : UT to 24: UT, n the omprison etween fof2() n fof2(q) mesure t the hosen sttion is rrie out if the time ifferene etween fof2() n fof2(q) mesurements is less thn or equl to 3 ys. We use the nerest quiet y to the stuie isture time perio, n etermine the reltive evition, δ, of fof2 oserve t the given sttion from fof2(q) s δ = fof2()/fof2(q) 1 = (NmF2()/NmF2(q)) 1/2 1. (3) Negtive n positive vlues of δ orrespon to negtive n positive isturnes in NmF2, respetively. We stuy the epenene of the proilities of the negtive isturne ourrenes in NmF2 on χ. Following the preeing work y Lozin n Pvlov (22), we give negtive fof2 isturnes the lels wek (.1 < δ < or.81<nmf2()/nmf2(q)<1), norml (.3 < δ.1 or.49<nmf2()/nmf2(q).81), strong (.5 < δ.3 or.25<nmf2()/nmf2(q).49), n very strong (δ.5 or NmF2()/NmF2(q).25), n onfine our ttention to reltionships etween them n the G onition or F1-lyer ourrenes. Similr to the G (χ) n F 1 (χ) etermintions, we nlyze the mesure fof1 n fof2 within eh oveetermine ltitue, month n lol time rnge. We etermine the proility, δ1 δ δ (χ), of the NmF2 negtive isturne ourrene s rtio of numer of NmF2 negtive isturne oservtions within the δ1 δ δ rnge for zenith ngles within the χ intervl to totl numer of stuie NmF2 negtive n positive isturne oservtions for the sme χ, within the given ltitue, month, n lol time rnge. 4 Results n isussion 4.1 F1-lyer n G onition ourrene proilities Similr to the preeing work y Lozin n Pvlov (22), the totl numer of hourly mesurements stuie is whih inlues G onition ourrenes n F1-lyer ourrenes. Our NmF2 isturne nlysis inlues only negtive n positive ionospheri isturnes tht hve referene quiet ys (see Set. 3). A prt

6 1826 V. V. Lozin n A. V. Pvlov: Solr zenith ngle epenenies 3 :-12: LT 12:-24: LT.4 :-12: LT 12:-24: LT.3 Ψ F1 2 Ψ G Fig. 1. The epenene of the F1-lyer perentge ourrenes on the solr zenith ngle in ltitue rnge 1 ( 1 ) uring the winter (soli lines), summer (she lines), n spring n utumn Fig. 1 (otte lines) months. The 18 solr zenith ngle rnge inlues the lol time perio from : LT to 12: LT (left pnel) n from 12: LT to 24: LT (right pnel). Ψ F :-12: LT 12:-24: LT of the hourly fof2 isturne mesurements hs no referene quiet ys, in greement with the quiet y efinition epte in our pper, n these hourly fof2 mesurements re not nlyze. Averge vlues of F 1 (χ) n G (χ) re presente in Tle 1. For eh stuie ltitue rnge n seson, the first numer is etermine s n verge vlue, < F 1 (χ) > 1, of F 1 (χ) or n verge vlue, < G (χ) > 1, of G (χ) for the first hlf of y for χ 9, while the seon numer is etermine s n verge vlue, < F 1 (χ) > 2, of F 1 (χ) or n verge vlue, < G (χ) > 2, of G (χ) for the seon hlf of y for χ 9. Tle 1 shows tht < F 1 (χ) > 2 is less thn < F 1 (χ) > 1, exept for ltitue rnge 5 uring the winter, spring, n utumn months. Our lultions show tht < G (χ) > 2 is less thn < G (χ) > 1 in ltitue rnges 2 5 for ll sesons, exept for ltitue rnge 3 in winter. For ltitue rnge 3, < G (χ) > 2 is pproximtely equl to < G (χ) > 1 for ll sesons. The epenenies of the F1-lyer perentge ourrenes on the solr zenith ngle in ltitue rnge 1 re shown in Fig. 1 for the lol time perio from : LT to 12: LT (left pnel) n from 12: LT to 24: LT (right pnel). Figures 2 5 show the epenene of the F1-lyer (ottom pnels) n G onition (top pnels) perentge ourrene on the solr zenith ngle in ltitue rnge 2 (Fig. 2), in ltitue rnge 3 (Fig. 3), in ltitue rnge 4 (Fig. 4), n in ltitue rnge 5 (Fig. 5), uring the winter (soli lines), summer (she lines), n spring n utumn (otte lines) months. The left pnels of Figs. 2 5 represent the F1-lyer n G onition perentge ourrene from : LT to 12: LT, while the right pnels of Figs. 2 5 give the F1-lyer n G onition perentge ourrene from 12: LT to 24: LT. It shoul e note tht the lulte vlue of the G onition Fig. 2. The epenene of the F1-lyer (two ottom pnels), n G Fig. 2 onition (two top pnels) proility funtions on the solr zenith ngle in ltitue rnge 2 (1 < 3 ) uring the winter (soli lines), summer (she lines), n spring n utumn (otte lines) months efore miy from : LT to 12: LT (left pnels), n fter miy from 12: LT to 24: LT (right pnels). ourrene proility is negligile ove the geomgneti equtoril region (ltitue rnge 1) uring ll sesons n in ltitue rnge 2 in winter. Therefore, the G (χ) epenenies re not isusse in this work for these ses. Figures 2 5 show tht the G onition is more likely to our uring the first hlf of y thn uring the seon hlf of y in ltitue rnges 2 5 uring ll sesons for the sme vlue of the solr zenith ngle, exept for ltitue rnge 3 in winter, when the G onition ourrene proility is pproximtely the sme for the sme solr zenith ngle efore n fter 12: LT. The F1-lyer ourrene proility is lrger in the first hlf of y in omprison with tht in the seon hlf of y for the sme vlue of the solr zenith ngle within ltitue rnge 1 for ll sesons, while the F1- lyer ourrene proility is pproximtely the sme for the sme solr zenith ngle efore n fter 12: LT in ltitue rnges 4 n 5. It n e seen from Figs. 1 5 tht the F1-lyer n G onition re more ommonly forme ner miy thn ner post sunrise or pre-sunset, when the F-region is in the sunlight. These results re in greement with the onlusions

7 V. V. Lozin n A. V. Pvlov: Solr zenith ngle epenenies :-12: LT 12:-24: LT :-12: LT 12:-24: LT 2.8 Ψ G Ψ G.4 1 :-12: LT 12:-24: LT 6 :-12: LT 12:-24: LT 4 Ψ F1 2 Ψ F Fig. 3. The epenene of the F1-lyer (ottom pnels), n G onition (top pnels) proility funtions Fig. 3 on the solr zenith ngle in ltitue rnge 3 (3 < ϕ 45, 3 < 45 ) uring the winter (soli lines), summer (she lines), n spring n utumn 2 (otte lines) months efore miy from : LT to 12: LT (left pnels), n fter miy from 12: LT to 24: LT (right pnels). Fig. 4. The epenene of the F1-lyer (ottom pnels), n G onition (top pnels) proility funtions Fig. 4 on the solr zenith ngle in ltitue rnge 4 (45 < ϕ 6, 45 < 6 ) uring the winter (soli lines), summer (she lines), n spring n utumn (otte lines) months efore miy 3 from : LT to 12: LT (left pnels), n fter miy from 12: LT to 24: LT (right pnels). of previous stuies (for more etils, see, for exmple, Rtliffe, 1956, 1972; Yonezw et. l., 1959; Polykov et l., 1968) se on the theoretil stuies n the limite t set. Figures 1 5 provie for the first time the quntittive mesure of the proility vritions with solr zenith ngle hnges. Figures 1 4 show tht the mximum vlues of the F1-lyer n G onition ourrene proilities re lote in the 45 solr zenith ngle rnge in ltitue rnges 1 4. The mximum vlues of the F1-lyer n G onition ourrene proilities re relize for the minimum vlue of χ lose to noon in ltitue rnge 5 (see Fig. 5). At the minimum solr zenith ngle vlues (see Figs. 2 3), the numer of oservtions is lrge enough for the ientifile osilltions in the proilities t these solr zenith ngles to e sttistilly signifint. The physil resons for the ourene of these osilltions re unler. The omprison in the vlues of F 1 (χ) n G (χ) etween the five ltitue regions esrie ove shows the ytime teneny for erese in these proilities t low ltitues n n inrese in these proilities t high ltitues for ll sesons. In the previous F1-lyer n G onition stuies (Rtliffe, 1956, 1972; Polykov et l., 1968; Shhepkin et l., 1984) se on the limite t set, it ws emonstrte tht the hne tht the F1-lyer n G onition will e forme is greter in summer thn in winter. Lozin n Pvlov (22) hve provie itionl eviene of this phenomenon y lulting for the first time the iurnlly verge F1-lyer n G onition proility vritions with seson. Comprison etween soli (winter months), she (summer months), n otte (spring n utumn months) lines in Figs. 2 5 gives more etile piture of the F1- lyer n G onition sesonl proility ehvior for the given solr zenith ngle in eh ltitue rnges 1 5, therey proviing eviene tht the hne tht the ytime F1-lyer n G onition will e forme is greter in summer thn in winter. We hve foun for the first time tht the F1-lyer ourrene proility is greter in winter thn in summer for ll solr zenith ngles ove the geomgneti equtoril region (see Fig. 1). It shoul lso e note tht the F1- lyer n G onition sesonl proilities re lower uring the spring n utumn months s ompre with tht uring

8 1828 V. V. Lozin n A. V. Pvlov: Solr zenith ngle epenenies Ψ G :-24: LT 12:-24: LT rnges 1 3 with the F1-lyer ourrene proility shown in Figs It is lso unler why there re no sesonl ifferenes in the epenene of P S on χ. The etils of eriving Eq. (4) re not presente in the short pper y Sotto et l. (1998), n we nnot give n explntion of the ientifile ifferenes etween our results n P S (, χ). 4.2 NmF2 norml, strong n very strong negtive isturne ourrene proilities Ψ F :-12: LT 12:-24: LT Fig. 5 Fig. 5. The epenene of the F1-lyer (ottom pnels), n G onition (top pnels) proility funtions on the solr zenith ngle in ltitue rnge 5 (6 < 9 ) uring the winter (soli lines), summer (she lines), n spring n utumn (otte lines) 4 months efore miy from : LT to 12: LT (left pnels), n fter miy from 12: LT to 24: LT (right pnels). the summer months for most of the solr zenith ngle rnge in ltitue rnges 3 5 (ompre otte n she lines of Figs. 3 5). Sotto et l. (1998) foun tht the proility of evluting the ourrene of the F1-lyer n e presente s P S (, χ) = 1 (1 (χ/9)) K for χ < 9, n P S (, χ) = for χ 9, (4) where K = , n the unit of P S is perent. The nlysis of Sotto et l. (1998) ws se on t quire y the Ionospheri Digitl Dtse of the Ntionl Geophysil Dt Center, Bouler, Coloro, from 1969 to 199. The vlue of P S (, χ) pprohes 1% for χ = n % for χ = 9. By ompring the epenene of the F1-lyer ourrene proility on the solr zenith ngle foun in our work in eh ltitue rnge with P S (, χ), we onlue tht P S (, χ) overestimtes the rel vlue of the F1-lyer ourrene proility shown in Figs For exmple, 8% P S (, χ) 1% for χ 52 in the geomgneti ltitue rnge of 45 45, whih inlues ltitue Averge vlues of ll stuie NmF2 negtive isturne proilities lulte in ll ltitue rnges n uring ll sesons re presente in Tle 2. For eh sort of NmF2 negtive isturne, stuie ltitue rnge n seson, the first numer is etermine s n verge vlue, < δ1 δ δ (χ) > 1 of δ1 δ δ (χ) for the first hlf of y for χ 9, while the seon numer is etermine s n verge vlue, < δ1 δ δ (χ) > 2, of δ1 δ δ (χ) for the seon hlf of y for χ 9. The thir numer is etermine s n verge vlue of δ1 δ δ (χ) for the night-time perio for χ > 9. An verge ytime vlue of ny NmF2 negtive isturne proility is lulte s hlf-sum of the first n seon numers given in Tle 2. Tle 2 shows tht < δ1 δ δ (χ) > 2 is less thn the < δ1 δ δ (χ) > 1 for the norml, strong or very strong NmF2 negtive isturnes in ltitue rnges 3 5 for ll sesons, exept for the very strong NmF2 negtive isturnes in ltitue rnges 3 in the winter n in ltitue rnge 5 uring winter, spring, n utumn months, when < δ.5 (χ) > 2 is pproximtely equl to < δ.5 (χ) > 1. In opposition to ltitue rnges 3 5, < δ.5 (χ) > 1 is less thn < δ.5 (χ) > 2 for the norml, strong, n very strong NmF2 negtive isturnes in ltitue rnges 1 n 2 in winter, exept for the very strong NmF2 negtive isturnes in ltitue rnge 2 uring ll sesons n tht in ltitue rnge 1 uring the summer n winter months. We foun tht < δ.5 (χ) > 1 is pproximtely equl to < δ.5 (χ) > 2 in ltitue rnge 2 uring ll sesons. Figures 6 1 show the epenene of the NmF2 negtive isturne perentge ourrene on the solr zenith ngle in ltitue rnge 1 (Fig. 6), in ltitue rnge 2 (Fig. 7), in ltitue rnge 3 (Fig. 8), in ltitue rnge 4 (Fig. 9), n in ltitue rnge 5 (Fig. 1) uring the winter (soli lines), summer (she lines), n spring n utumn (otte lines) months for the wek (pnels ), norml (pnels ), strong (pnels ), n very strong (pnels ) NmF2 negtive isturnes. Left pnels give the results for the lol time perio from : LT to 12: LT, while the right pnels present the results of lultions from 12: LT to 24: LT. The first thing to note from Figs. 6 1 is tht our results lerly pture the ltitue epenene in the NmF2 norml, strong n very strong negtive isturne proilities, reprouing the generl teneny for erese in these proilities t low ltitues n n inrese in proilities t high ltitues, in greement with previous onlusions of Lozin n Pvlov (22), who hve lulte the iurnlly n se-

9 V. V. Lozin n A. V. Pvlov: Solr zenith ngle epenenies 1829 Tle 2. Averge vlues of wek, norml, strong, n very strong NmF2 negtive isturne perentge ourrene in ltitue rnges 1 5 uring the winter, summer, n spring n utumn months for the first (first numer) n seon (seon numer) hlf of y for χ 9, n uring the night-time perio for χ > 9 (thir numer) Ltitue rnge 1 Ltitue rnge 2 Ltitue rnge 3 Ltitue rnge 4 Ltitue rnge 5 Winter 31.3, 3.8, , 27.9, , 31.5, , 34.1, , 27.4, 18. Wek Summer 33.1, 28.3, , 28., , 3.9, , 36.6, , 36.4, 22.2 Spring, 3.3, 29.5, , 3.4, , 32.8, , 32.2, , 28.6, 17.4 utumn Winter 1.7, 13., , 16.5, , 12.5, , 17.4, , 33.1, 22.3 Norml Summer 11.8, 17.8, , 23.7, , 27.2, , 27.2, , 33.8, 33.6 Spring, 9.3, 12.5, , 17.2, , 21.7, , 29.8, , 36.7, 25. utumn Winter.1,.2, , 1.7, 5.5.8,.4, , 2., , 9.2, 9.3 Strong Summer.4,.6, , 3.9, , 3.3, , 1.7, , 3.2, 12.6 Spring,.2,.3, , 1.9, 6. 4., 3.4, , 4.7,, , 7.5, 11.6 utumn Winter.2,.1,.9.1,.1,.7.1,.1,.2.3,.2,.6 1.1, 1.1, 2.5 Very Summer.4,.3, 1..2,.1,.7.2,.1,.2.3,.1, 1.1.2,.1, 1.6 strong Spring,.2,.4,.8.2,.1,.9.3,.2,.3.5,.2, 1.2.5,.5, 3.3 utumn son verge epenenies of the NmF2 negtive isturne proility funtions on the geomgneti ltitue for δ <, δ.1, δ.3, n δ.5. Tle 2 n Figs. 6 n 7 show tht the NmF2 norml, strong, n very strong negtive isturnes re more frequent on verge t night thn y y in ltitue rnges 1 n 2 for ll sesons, rehing their mximum n minimum ourrene proility vlues t night n y y, respetively. This onlusion is lso orret for ll other stuie ltitue regions uring the winter months (see Tle 2 n soli lines in Figs. 8 1), exept for the NmF2 norml n strong negtive isturnes in ltitue rnge 5 (see Figs. 1, ). Tle 2 n the she n otte lines in Figs. 9, n Figs. 1, show tht the verge night-time ourrene proility is lrger thn the verge ytime ourrene proility for the strong n very strong negtive isturnes in ltitue rnges 4 n 5 uring the summer, spring, n utumn months. The Joule heting of the thermosphere n e viewe s the fritionl heting proue in the thermosphere s the rpily onveting ions ollie with neutrl moleules. Most of the Joule heting is eposite in the km ltitue region, lthough some extens to higher ltitues (Rihmon n Lu, 2). The geomgneti storm Joule heting of the thermosphere is onsierly more effetive thn the energy of the urorl eletrons in ffeting the thermospheri irultion n in the inrese in the neutrl temperture (Rihmon n Lu, 2). Joule heting from the issiption of ionospheri urrents rises the neutrl temperture of the upper thermosphere, n the ion rg rives the high-veloity neutrl wins uring geomgneti storms t high ltitues (Prölss, 198, 1995; Fuller-Rowell et l., 1996, 2). It les to genertion of isture omposition zone of the high-ltitue neutrl tmosphere, with n inrese in hevier gses n erese in lighter gses, i.e. with n inrese in [N 2 ] n [O 2 ] n in the [N 2 ]/[O] n [O 2 ]/[O] rtios. The win surge propgtes from uror regions to low ltitues in oth hemispheres. As result, thermospheri ltitue istriutions of neutrl speies t mile n low ltitues re influene y glol, lrge-sle win irultion whih is proue y geomgneti storm energy input t high ltitues (theoretil n oservtionl stuies of thermospheri omposition responses to the trnsport of neutrl speies from urorl regions to mile ltitues uring geomgneti storms re reviewe y Prölss, 198, 1995). The inrese in the [N 2 ]/[O] rtio mximises in region tht is roughly lote in the viinity of the urorl ovl, n this [N 2 ]/[O] inrese intensifies n n expn to mile mgneti ltitues with the K p inrese (Brunelli n Nmglze, 1988; Prölss, 198, 1995; Zuzi et l., 1997; Buonsnto, 1999). The high-ltitue geomgneti storm upwelling rings ir rih in the hevy speies N2 n O2 to high ltitues, n the geomgneti storm irultion rries this N 2 n O 2 -rih ir to mi-ltitues n lower ltitues. The geomgneti storm ownwelling les to the opposite effet: ir with low vlues of [N 2 ] n [O 2 ] is rrie ownwr, reuing their onentrtions t ll ltitues (e.g. Fuller-Rowell et l., 1996; Fiel et l., 1998; Rihmon n Lu, 2). Thus, the vlues of [N 2 ] n [O 2 ], n the [N 2 ]/[O] n the [O 2 ]/[O] rtios re more enhne t high ltitues thn t mile ltitues, ontriuting to more NmF2 ereses t high ltitues thn t mile ltitues. The geomgneti storm N 2 n O 2 numer ensities, n the [N 2 ]/[O] n the [O 2 ]/[O] rtios re eplete t low ltitues, using NmF2 inreses t low ltitues. As result, the ytime n night-time ltitue trens in the proilities of NmF2 negtive isturnes shown in Figs. 6 1 n rise from these ltitue trens in [N 2 ], [O 2 ], [N 2 ]/[O], n [O 2 ]/[O]. As ws esrie ove, n itionl enhne equtorwr win rises uring geomgneti storm, leing to

10 183 V. V. Lozin n A. V. Pvlov: Solr zenith ngle epenenies Ψ.1<δ< 3 2 :-12: LT 12:-24: LT Ψ.1<δ< 3 2 :-12: LT 12:-24: LT Ψ.3<δ Ψ.3<δ Ψ.5<δ Ψ.5<δ Ψ δ Ψ δ Fig. 6 Fig. 6. The epenene of the NmF2 negtive isturne proility funtions on the solr zenith ngle in ltitue rnge 1 ( 1 ) uring the winter (soli lines), summer (she lines), n spring n utumn (otte lines) 5 months for the wek (pnels ), norml (pnels ), strong (pnels ), n very strong (pnels ) NmF2 negtive isturnes. The 18 solr zenith ngle rnge inlues the lol time perio from : LT to 12: LT (left pnels) n from 12: LT to 24: LT (right pnels) Fig. 7 Fig. 7. The epenene of the NmF2 negtive isturne proility funtions on the solr zenith ngle in ltitue rnge 2 (1 < 3 ) uring the winter (soli lines), summer (she lines), n spring n utumn (otte lines) months for the wek (pnels 6 ), norml (pnels ), strong (pnels ), n very strong (pnels ) NmF2 negtive isturnes. The 18 solr zenith ngle rnge inlues the lol time perio from : LT to 12: LT (left pnels) n from 12: LT to 24: LT (right pnels). trnsport of neutrl omposition hnges to lower ltitues. The resulting equtorwr win is stronger t night, euse the itionl geomgneti storm equtorwr win is e to the quiet y-to-night irultion n euse the itionl win is reinfore y ntisunwr ion rg ue to mgnetospheri onvetion E B rifts (Strus n Shulz, 1976; Bok n Evns, 1979). As result, the neutrl omposition isturne zone rehes more lower ltitues t night thn y y, n the NmF2 norml, strong n very strong negtive isturnes ten to e more frequent on verge t night thn y y in ltitue rnges 1 n 2 for ll sesons (see Figs. 6 n 7). On the other hn, rise in hmf2 to regions with reue loss rte of O + ( 4 S) ions ue to equtorwr wins proues n inrese in NmF2, while rop in hmf2 ue to polewr wins reues NmF2. This ompetition etween neutrl omposition isturne using NmF2 negtive storm effets n rise in hmf2 using NmF2 positive storm effets etermines the omplite epenene of the norml, strong n very strong NmF2 neg- tive isturne perentge ourrenes on the solr zenith ngle (see Figs. 6 1). We foun tht there is ifferene etween the epenene of the strong n very strong NmF2 negtive isturne perentge ourrenes on the solr zenith ngle in ltitue rnges 1 n 2 (see Figs. 6, n Figs. 7, ). There re ler isriminte peks in the proilities of the strong n very strong NmF2 isturnes efore sunrise for ll sesons in ltitue rnge 1, while the strong n very strong NmF2 negtive isturne ourrene proilities re erese (with some osilltions) with the erese in the solr zenith ngle in the post minight night-time setor in ltitue rnge 2. Our lultions show (see Figs. 9, n Figs. 1, ) tht the.5<δ.3 (χ) pek n the δ.5 (χ) pek in the post minight night-time setor re ompnie y the.5<δ.3 (χ) pek n the δ.5 (χ) pek in the sunset-to-minight setor in ltitue rnges 4 n 5 uring the winter, spring, n utumn months.

11 V. V. Lozin n A. V. Pvlov: Solr zenith ngle epenenies 1831 :-12: LT 12:-24: LT 4 :-12: LT 12:-24: LT Ψ.1<δ< Ψ.1<δ< 3 2 Ψ.3<δ Ψ.3<δ Ψ.5<δ Ψ.5<δ Ψ δ Ψ δ Fig. 8 Fig. 8. The epenene of the NmF2 negtive isturne proility funtions on the solr zenith ngle in ltitue rnge 3 (3 < ϕ 45, 3 < 45 ) uring the winter (soli lines), summer (she lines), n spring n utumn (otte lines) months for 7 the wek (pnels ), norml (pnels ), strong (pnels ), n very strong (pnels ) NmF2 negtive isturnes. The 18 solr zenith ngle rnge inlues the lol time perio from : LT to 12: LT (left pnels) n from 12: LT to 24: LT (right pnels) Fig. 9 Fig. 9. The epenene of the NmF2 negtive isturne proility funtions on the solr zenith ngle in ltitue rnge 4 (45 < ϕ 6, 45 < 6 ) uring the winter (soli lines), summer (she lines), n spring n 8 utumn (otte lines) months for the wek (pnels ), norml (pnels ), strong (pnels ), n very strong (pnels ) NmF2 negtive isturnes. The 18 solr zenith ngle rnge inlues the lol time perio from : LT to 12: LT (left pnels) n from 12: LT to 24: LT (right pnels). 4.3 NmF2 wek negtive isturne ourrene proilities Averge vlues of the wek NmF2 negtive isturne proilities lulte for ll ltitue rnges n sesons re presente in Tle 2. For eh stuie ltitue rnge n seson, the first numer is etermine s n verge vlue, <.1<δ< (χ) > 1, of <.1<δ< (χ) for the first hlf of y for χ 9, while the seon numer is etermine s n verge vlue, <.1<δ< (χ) > 2, of <.1<δ< (χ) for the seon hlf of y for χ 9. The thir numer is etermine s n verge vlue of.1<δ< (χ) for the night-time perio for χ > 9. An verge ytime vlue of the wek NmF2 negtive isturne proility is lulte s hlf-sum of the first n seon numers given in Tle 2, shows tht <.1<δ< (χ) > 2 is less thn <.1<δ< (χ) > 1 in ltitue rnges 3 5 for ll sesons, exept for ltitue rnge 3 in winter n ltitue rnge 5 for the spring n utumn months. In opposition to ltitue rnges 3 5, where <.1<δ< (χ) > 1 is less thn <.1<δ< (χ) > 2 in ltitue rnges 1 n 2 for ll sesons, exept for ltitue rnge 2 for spring n utumn months. The top pnels of Figs. 6 1 show tht the ourrene proility of the NmF2 wek negtive isturnes rehes its mximum n minimum vlues uring ytime n nighttime onitions, respetively. It lso follows from Tle 2 tht the verge night-time vlue of this proility is less thn tht y y for ll sesons in ll ltitue regions. It shoul e note tht the F2-lyer retion to wek storms n to sustorms is not esily iretly oservle, sine wek effets re mske y ioniztion vritions or y other vritions tht re not relte to geomgneti tivity (hour-to-hour n y-to-y vriility, et.). The vritions in the neutrl tmosphere, the neutrl wins, n the solr EUV flux n e reflete in the vriility of NmF2, n it is sertine tht night-time vriility is lrger thn ytime vriility of NmF2 (Fores et l., 2). Pulishe

12 1832 V. V. Lozin n A. V. Pvlov: Solr zenith ngle epenenies Ψ.1<δ< Ψ.3<δ.1 Ψ.5<δ.3 Ψ δ :-12: LT 12:-24: LT Fig. 1 Fig. 1. The epenene of the NmF2 negtive isturne proility funtions on the solr zenith ngle in ltitue rnge 5 (6 < 9 ) uring the winter (soli lines), summer (she lines), n spring n utumn (otte lines) months for the wek (pnels 9 ), norml (pnels ), strong (pnels ), n very strong (pnels ) NmF2 negtive isturnes. The 18 solr zenith ngle rnge inlues the lol time perio from : LT to 12: LT (left pnels) n from 12: LT to 24: LT (right pnels). vlues of ionospheri eletron ontent were use y Arvinkshn n Iyer (1993) to stuy its y-to-y vriility t numer of sttions extening from equtoril to miltitues in Inin n Amerin setors for high n low solr tivity yers. The vriility is lrger t night thn y y, highest in Ferury n Novemer n lowest in equinox months (Arvinkshn n Iyer, 1993). As result, we onlue tht the ientifile greter proility of the NmF2 wek negtive isturnes y y thn t night (see the top pnels of Figs. 6 1) is not relte with vriility in the ionosphere. In ition to the moifie lrge-sle irultion of the neutrl tmosphere, uring geomgneti isturnes, the sptil n temporl vritions of high-ltitue thermosphere het soures exite lrge mplitue grvity wves, whih proue trvelling ionospheri isturnes in the F-region of the ionosphere (Millwr et l., 1993; Hoke n Shlegel, 1996). Suh grvity wves propgte from high to low ltitues onsierly fster in the thermosphere thn typil mi- n low-ltitue wins resulting from storms (Rees, 1995; Hoke n Shlegel, 1996). The response of the miltitue ionosphere to this grvity wves propgtion is oserve y ionozone sttions n inoherent stter rrs in the rising or lowering of hmf2, often y severl 1s of km, leing to erese or n inrese in L n [O] t hmf2, i.e. leing to the inrese or erese in NmF2, respetively (Rees, 1995; Hoke n Shlegel, 1996). The nlysis of fof2 mesurements shows tht night-time fof2 ereses ue to grvity wve propgtion re not so signifint s y y (Deminov et l., 1998). As result, we onlue tht the ientifile greter proility of the NmF2 wek negtive isturnes y y thn t night, shown in the top pnels of Figs. 6 1, n e expline if we suggest tht NmF2 wek negtive isturnes re rete y grvity wve propgtion in the ionosphere. 4.4 Reltionships etween the F1-lyer, NmF2 negtive isturne, n G onition ourrene proility epenenies on χ Figures 2 5 show tht the ytime epenene of the F1- lyer ourrene proility on the solr zenith ngle is generlly in phse with tht for the G onition in ltitue rnges 3 5 for ll sesons, n in ltitue rnge 2 uring the spring, summer, n utumn months. However, we n onlue from Fig. 2 tht this ytime oupling is less onvining in ltitue rnge 2 uring the winter months. The ompetition etween the F1 n F2-lyers for ensity ominne etermines the G onition ourrene proility. Therefore, the ourrene proilities of the wek, norml, strong, n very strong NmF2 negtive isturnes, in ition to the F1-lyer ourrene proility, must e onsiere in ressing the uses of G onition solr zenith ngle hnges n in stuying the possile reltionships etween the F1- lyer n NmF2 negtive isturne ourrenes. The G onition in the geomgnetilly isture ionosphere is ssoite minly with signifint negtive ionospheri storm in NmF2 (Lozin n Pvlov, 22). Thus, the foun F1- lyer ourrene proility epenene on the solr zenith ngle (see the low pnels of Figs. 2 5) n the ientifile solr zenith ngle trens in strong n very strong negtive isturne proilities shown in pnels () n () of Figs. 7 1, re the trens involve in the formtion of the G onition solr zenith ngle tenenies shown in the top pnels of Figs A erese in the solr zenith ngle les to ereses in the ytime vlues of strong n very strong negtive isturne ourrene proilities in ltitue rnges 4 n 5 for the spring, summer, n utumn months (she n otte lines in pnels () n () of Figs. 9 n 1). The ytime proilities.3 δ<.1 (χ), n δ.5 (χ) of the NmF2 negtive isturne ourrene o not show isriminte trens in n inrese or erese with the solr zenith ngle erese in ltitue rnge 2 for ll sesons (pnels () n () of Fig. 7), n in ll the stuie ltitue regions for the winter months (soli lines in pnels () n () of Figs. 6 1). This