Vol.: (0123456789)
J Seismol (2025) 29:147–169
h ps://doi.o g/10.1007/s10950-024-10264-x
RESEARCH
Seismic scena io simula ion andANN‑based g ound mo ion
model de elopmen on heNo h Tab iz Faul inNo hwes
I an
Cagla Temiz· S.M.SajadHussaini· ShaghayeghKa imzadeh·
AysegulAskan· PauloB.Lou enço
Recei ed: 23 July 2024 / Accep ed: 25 Oc obe 2024 / Published online: 20 No embe 2024
© The Au ho (s) 2024
Abs ac Ea hquakes pose signi ican seismic haz-
a ds in u ban egions, o en causing ex ensi e damage
o he buil en i onmen . In egions lacking obus
seismic moni o ing ne wo ks o su icien da a om
his o ical e en s, g ound mo ion simula ions a e c u-
cial o assessing po en ial ea hquake impac s. Ye ,
alida ing hese simula ions is challenging, leading
o no able p edic i e unce ain y. This s udy aims
o simula e ou scena io ea hquakes wi h momen
magni udes o 6.8, 7.1, 7.4, and 7.7 in I an, speci i-
cally in es iga ing a ia ions in aul plane up u e
and ea hquake hypocen e . The No h Tab iz Faul
(NTF), loca ed wi hin he seismic gap in no hwes
I an, is selec ed as he case s udy due o he lack o
well- eco ded g ound mo ions om se e e ea h-
quakes, despi e his o ical e idence o la ge-mag-
ni ude e en s. Simula ions a e conduc ed using a
s ochas ic ini e- aul g ound mo ion simula ion me h-
odology wi h a dynamic co ne equency. Valida-
ion o he simula ions is pe o med by compa ing
es ima ed peak g ound mo ions and pseudo-spec-
al o dina es wi h exis ing g ound mo ion models
(GMMs), supplemen ed by in e -pe iod co ela ion
analysis. Simula ion esul s e eal high haza d le -
els, especially in he no heas e n a ea nea he aul
plane. In ensi y maps in e ms o he Modi ied Me -
calli In ensi y (MMI) scale unde sco e he u gency
o comp ehensi e p epa edness measu es. Finally,
a egion-speci ic GMM is de eloped using A i i-
cial Neu al Ne wo ks (ANN) o p edic peak g ound
mo ion pa ame e s wi h an online pla o m accessible
o end-use s.
Keywo ds S ochas ic g ound mo ion simula ion·
Seismic haza d maps· Modi ied me calli in ensi y
(MMI) map· No h Tab iz Faul (Tab iz I an)·
A i icial neu al ne wo ks (ANN)-based g ound
mo ion model (GMM)
1 In oduc ion
F om ancien imes o he p esen day, ea hquakes
ha e emained one o he mos de as a ing na u al
C.Temiz
Depa men s o Ci il Enginee ing, Middle Eas Technical
Uni e si y, Anka a, Tü kiye
e-mail: cagla [email p o ec ed]
S.M.S.Hussaini· S.Ka imzadeh(*)· P.B.Lou enço
Ci il Enginee ing Depa men , Ins i u e o Sus ainabili y
andInno a ion inS uc u al Enginee ing (ISISE), ARISE,
Uni e si y o Minho, Guima aes, Po ugal
e-mail: [email p o ec ed]
S. M. S. Hussaini
e-mail: [email p o ec ed]
P. B. Lou enço
e-mail: [email p o ec ed]
A.Askan
Depa men s o Ci il Enginee ing andEa hquake S udies,
Middle Eas Technical Uni e si y, Anka a, Tü kiye
e-mail: [email p o ec ed]
148
J Seismol (2025) 29:147–169
Vol:. (1234567890)
disas e s, leading o signi ican casual ies and he
des uc ion o in as uc u es in seismically ac i e
a eas. Howe e , ecen ad ancemen s in echnology
and compu a ional powe ha e imp o ed he abili y
o assess po en ial ea hquake haza ds, enabling he
design o sa e s uc u es ha help mi iga e he e ec s
o ea hquakes. In egions wi h well-es ablished seis-
mic ne wo ks ha a e p one o equen and se e e
ea hquakes, eco ded g ound mo ions can be used
as inpu s o p ecise isk assessmen s udies, he eby
imp o ing he sa e y o he buil en i onmen . How-
e e , in a eas wi h seismic gaps and limi ed eco ded
g ound mo ions, pa icula ly om s ong ea hquakes
a sho sou ce- o-si e dis ances, he a ailabili y o
egion-speci ic g ound mo ion da a is o en insu i-
cien o de ining a design scena io (Rezaeian e al.
2024; Yamamo o and Bake 2013). Addi ionally, he
p ac ice o selec ing and scaling g ound mo ions can
al e he co ela ion be ween he modi ied mo ions
and hei o iginal physical condi ions (Hancock e al.
2006; Wa son-Lamp ey and Ab ahamson 2006). In
such cases, g ound mo ion simula ions o e an al e -
na i e app oach o gene a ing ep esen a i e ime
se ies ailo ed o he egion’s seismological ea u es
(A o a e al. 2020). These simula ions can be used o
e alua e he an icipa ed haza d le els and he esul -
ing po en ial seismic damage and isk le els in seis-
mic gap egions wo ldwide.
The li e a u e ou lines a ange o g ound mo ion
simula ion echniques, each wi h a ying le els o
accu acy and equi ing di e en inpu -model pa am-
e e s and compu a ional app oaches. These ech-
niques can be b oadly ca ego ized in o de e minis ic,
s ochas ic, and hyb id me hods (Rezaeian and Sun
2014). De e minis ic me hods use ep esen a i e seis-
mic sou ce up u e and eloci y models alongside
nume ical me hods o sol e he pa ial di e en ial
equa ions go e ning wa e p opaga ion (F ankel 1993).
These me hods a e mos e ec i e o simula ing low-
equency con en s in egions lacking high- esolu ion
ea h models due o he minimum wa eleng h con-
s ain s. In con as , s ochas ic me hods a e be e
sui ed o simula ing highe - equency con en ha
is andom and incohe en . In he S ochas ic me hod,
a a - ield S-wa e spec um is enhanced by andom
phases o gene a e a mean ho izon al g ound mo ion
componen (Da id M. Boo e 1983). Hyb id me hods
combine bo h de e minis ic and s ochas ic app oaches
o accu a ely simula e he ull equency spec um and
le e age hei bene i s (Kamae e al. 1998).
G ound mo ion simula ions p o ide aluable
insigh s in o he po en ial impac s o ea hquakes,
pa icula ly in egions wi h limi ed eco ded seismic
ac i i y, enabling mo e accu a e isk assessmen s
and he de elopmen o a ge ed mi iga ion s a e-
gies. G ound mo ion simula ions ha e been widely
used in a ious s udies o eplica e his o ical e en s,
e alua e di e en ea hquake scena ios, assess s uc-
u al demands ac oss a ious ypes o s uc u es, and
de elop g ound mo ion models (GMMs).
Wi hin his con ex , Ka imzadeh and Askan (2018)
simula ed he his o ical 1939 E zincan ea hquake
in Tü kiye, compa ing hei esul s o empi ical
in ensi y maps. Simila ly, Tanı can and Yelkenci-
Necmioğlu (2020) simula ed he 2017 Bod um-Kos
ea hquake, alida ing hei esul s h ough compa i-
son wi h empi ical GMMs and gene a ing seismic
haza d maps. La e , Can e al. (2021) simula ed he
2002 Cay ea hquake in Tü kiye and alida ed hei
indings by compa ing goodness-o - i sco es be ween
obse ed and simula ed mo ions in addi ion o com-
pa ing hei esul s o empi ical in ensi y maps. These
s udies unde sco e he eliabili y o simula ions,
when p ope ly alida ed, in gene a ing obus seismic
haza d es ima es compa able o empi ical GMMs and
obse ed in ensi ies.
In addi ion o his o ical e en modeling, la ge
sui es o simula ions ha e con ibu ed o he de el-
opmen o GMMs, suppo ing he es ima ion o peak
g ound accele a ion (PGA) and spec al o dina es
(Ka imzadeh e al. 2023a; Ka imzadeh e al. 2023b;
Wi he s e al. 2020). Simula ions ha e also been used
o es ima e seismic demands o di e en s uc u al
sys ems, anging om single-deg ee-o - eedom sys-
ems (A kinson e al. 2011) o mo e complex ein-
o ced conc e e and s eel s uc u es (Ozsa ac e al.
2021; Tsioulou e al. 2019; Xu and Feng 2018).
Building on hese de elopmen s, Sme zini e al.
(2024) de eloped he BB-SPEEDse , a da ase o
nea -sou ce, b oadband physic-based simula ed
accele og ams, o b idge knowledge gaps caused by
spa se g ound mo ion eco dings. They alida ed
he simula ions om bo h seismological and engi-
nee ing pe spec i es o ensu e eliabili y o s uc-
u al assessmen s. In a ela ed s udy, Ha i i-A debili
and Rezaeian (2024) explo ed he applicabili y o
149
J Seismol (2025) 29:147–169
Vol.: (0123456789)
s ochas ic g ound mo ion simula ions o p obabilis ic
seismic pe o mance assessmen s o geo-s uc u es,
such as conc e e dams, by de eloping agili y cu es
and iden i ying op imal in ensi y measu es. Mo e
ecen ly, Ka imzadeh e al. (2024) simula ed he 1998
Faial ea hquake in he Azo es, ocusing on he s uc-
u al demands o mason y buildings, which a e o en
ulne able o he high- equency componen s o seis-
mic mo ions.
Recen e o s ha e emphasized he impo ance o
egion-speci ic simula ions o enhance p e-ea hquake
isk assessmen s, especially in seismically ac i e a eas.
Fo ins ance, A slan Kelam e al. (2022) and Askan
e al. (2024) applied g ound mo ion simula ions o e al-
ua e seismic isks in ulne able egions, emphasizing
he impo ance o scena io modeling o haza d iden-
i ica ion and mi iga ion s a egies. The signi icance o
such p oac i e assessmen s was highligh ed by he Feb-
ua y 2023 ea hquakes in Kah amanma as (Mw = 7.8)
and Elbis an (Mw = 7.5) in Tü kiye, which esul ed in
de as a ing casual ies, unde sco ing he need o e alu-
a e seismic isks be o e majo e en s occu .
Collec i ely, hese s udies showcase he g owing
impo ance o g ound mo ion simula ions in seismic
haza d and isk assessmen s, as well as in analyz-
ing seismic demands o di e en s uc u al sys ems,
assuming p ope calib a ion. Howe e , he p ecision
o hese assessmen s la gely depends on he selec ed
simula ion me hodologies and he quali y o hei cal-
ib a ion and alida ion p ocesses.
This s udy aims o simula e g ound mo ions in an
impo an seismic gap loca ed on he No h Tab iz
Faul (NTF) in No hwes I an while conside ing
he unce ain y ela ed o up u e e ec s. Tab iz is
selec ed as he s udy a ea due o i s designa ion as
one o he wo ld’s haza dous egions wi h a seismic
gap. The NTF, cha ac e ized by a sha p ace wi h a
No h-Wes o Sou h-Eas end, a e ses he no h-
e n pa o he ci y cen e . Ghasemi e al. (2021) ha e
examined he an icipa ed le els o damage esul ing
om he up u e o wo scena io e en s wi h momen
magni udes (Mw) o 7.4 and 7.3. P e ious s ud-
ies ha e sugges ed ha he NTF has he po en ial o
expe ience an ea hquake wi h a maximum Mw o
7.7 (Solaymani Azad e al. 2015). Despi e his o i-
cal ea hquakes ha ha e caused signi ican loss o
li e and widesp ead des uc ion, he e is a lack o
eco ded s ong ea hquakes in ecen his o y. This
highligh s he need o e alua e he po en ial haza d
le el o o hcoming seismic e en s. Al hough se -
e al s udies ha e ocused on seismic haza d, ulne -
abili y, and isk assessmen in he egion (Eslami
and Taghizadeh-Fa ahmand 2020; Ka imzadeh e al.
2014; Mohammadi e al. 2020a, b), he e emains a
signi ican gap in he in es iga ion o la ge-magni ude
ea hquake scena io simula ions speci ic o his a ea
and he de elopmen o egion-speci ic GMM.
This s udy pa icula ly add esses a unique p ob-
lem by in es iga ing he alea o y unce ain y associ-
a ed wi h he aul plane up u e and he loca ion o
he hypocen e on he NTF. Th ough g ound mo ion
simula ion, and alida ion agains a ious empi i-
cal g ound mo ion models, he seismic haza d in he
egion is also in es iga ed. To es ima e he seismic
haza d le el in he ci y cen e , we use he s ochas ic
ini e- aul g ound mo ion simula ion me hod wi h
a dynamic co ne equency app oach p oposed by
(Mo azedian and A kinson 2005). To model he mos
ca as ophic ea hquake scena ios, simula ions a e
conduc ed o po en ial ea hquakes wi h a ying
magni udes, including Mw = 7.7, 7.4, 7.1, and 6.8.
The simula ion esul s a e p esen ed and analyzed
h ough bo h isual and s a is ical me hods. Ini ially,
o alida e he simula ions, ou g ound mo ion mod-
els (GMMs) a e used, deno ed as BA08 (Boo e and
A kinson 2008), AC10 (Akka and Cagnan 2010),
ASB14 (Akka e al. 2014), and KAAH15 (Kale e al.
2015). BA08 is a globally applicable GMM, while
AC10 and ASB14 ocus on he Middle Eas e n and
Eu opean egions, and KAAH15 speci ically in ol es
da a om Eas e n Tü kiye and I an. Fo u he ali-
da ions o he simula ed da ase , his s udy del es
in o he in e -pe iod co ela ion o he esiduals o
pseudo-spec al o dina es. Following he alida ion
s eps, an icipa ed haza d measu es a e p esen ed
using de e minis ic seismic haza d maps ha display
a ious scena io e en s and hei associa ed unce -
ain ies. Addi ionally, a mean in ensi y map based on
he Modi ied Me calli In ensi y (MMI) scale is gen-
e a ed. Finally, a egional a i icial neu al ne wo k
(ANN)-based GMM is de eloped o p edic peak
g ound mo ion pa ame e s o an icipa ed scena ios in
he egion. Ul ima ely, he s udy p oduces an ex en-
si e compila ion o g ound mo ion ime se ies along
wi h he ANN-based GMM in ended o u u e appli-
ca ions in enginee ing p ac ice in an online pla o m.
150
J Seismol (2025) 29:147–169
Vol:. (1234567890)
2 S udy a ea
Tab iz is a p ominen ci y in I an wi h a ich his o i-
cal he i age. The a chaeological exca a ions indica e a
his o y o u baniza ion spanning app oxima ely 5000
yea s. P esen ly, Tab iz se es as he capi al o Eas -
Aze baijan P o ince and is si ua ed in he no hwes
egion o I an. Geog aphically, Tab iz lies be ween
eas e n Ana olia, he wes e n Caspian Sea, he sou he n
Caucasus h us bel , and he no he n Zag os Moun-
ains ange. Acco ding o he S a is ical Cen e o I an
(2020), he popula ion o Tab iz sligh ly exceeds 3 mil-
lion. The NTF is a p ominen geological ea u e in he
egion, consis ing o no hwes e n and sou heas e n
segmen s ha s e ch om Mishu Moun ain in he wes
o Bos anabad in he eas (Fa ahani 2018). This aul
has an a e age s ike o 115°, a leng h o app oxima ely
240 km, a dip o p edominan ly e ical, and a mini-
mum slip a e o 2 mm/yea (Ka akhanian e al. 2004).
The u ban a ea o Tab iz comp ises young and
unconsolida ed deposi s om i e s and glacial sedi-
men s o Cenozoic and Qua e na y o ma ions. Addi-
ionally, Tab iz is loca ed in an allu ial basin (Aza a za
and Ghazi a d 2016). The allu ial hickness in Tab iz
a ies by egion. In sou he n Tab iz, i eaches up o
250–300 m. Howe e , in he inhabi ed a eas, includ-
ing he ci y cen e ( he ac ual s udy a ea), he allu ial
hickness changes om 50 o 100 m (Ka imzadeh
e al. 2014). Mo eo e , exca a ions in he a ea ha e
e ealed he p esence o conglome a e, sands one, and
ma l laye s benea h hese allu ial sedimen s (Aza a za
and Ghazi a d 2016). The u banized a ea is mos ly on
clay and sil , and ma l and sands one laye s (Ka imza-
deh e al. 2014). The sou he n and eas e n pa s o he
ci y a e cha ac e ized by conglome a e and ma l laye s,
while sands one o ma ions a e ound in he no he n
and no heas e n moun ainous a eas.
Be be ian and A shadi (1976) ha e in es iga ed
he seismic ac i i ies o he NTF, including bo h his-
o ical e en s and hose in he wen ie h cen u y. P e-
ious s udies indica e ha Tab iz su e ed signi ican
de as a ion om mul iple ea hquakes be ween 634
A.D. and he la e 1800s. No ably, h ee majo ea h-
quakes in 1042 (Mw = 7.3), 1721 (Mw = 7.3), and
1780 (Mw = 7.4) caused se e e damage h ough co-
seismic su ace aul ing, as documen ed in epo s
om he nine een h cen u y and ea lie (Be be ian and
Yea s 1999; Hessami e al. 2003). Howe e , he exis -
ing desc ip ions o hese e en s lack su icien de ail o
accu a ely assess he ex en o des uc ion and g ound
de o ma ions (Be be ian and A shadi 1976). No ably,
Tab iz Ci y expe ienced subs an ial impac s om he
up u es along he no hwes e n segmen o he aul
in 1721 and 1780, as documen ed by K. Hessami e al.
(2009). While seismic ac i i y eco ds became mo e
eliable in he wen ie h cen u y, hese e en s did no
esul in majo damage o he ci y. Figu e1 p esen s he
egional ec onic map wi h he associa ed ea hquakes.
Gi en he long in e al since he las la ge-magni-
ude ea hquake (Mw ≥ 7.5) and he ongoing ac i i y
o he NTF, he po en ial o u u e s ong ea hquakes
in he a ea should no be igno ed. The e o e, he s udy
a ea is cen e ed on Tab iz ci y cen e . Khei izadeh
A ouq e al. (2020) conduc ed a s udy based on ca as-
ophe heo y, e ealing ha o e a qua e (35%) o
u ban a eas a e classi ied as highly ulne able o seis-
mic ac i i y. Addi ionally, Tab iz is iden i ied as a e y
highly suscep ible egion on he ea hquake suscep i-
bili y map de eloped by Mohammadi e al. (2020a)
This s udy simula es po en ial s ong ea hquakes and
ob ains simula ed g ound mo ions om 22 designa ed
dummy s a ions wi hin Tab iz ci y cen e , as illus a ed
in Fig.2. These s a ions a e selec ed based on high le els
o u baniza ion and human popula ion. The NTF ace is
shown wi h he ed line, while he blue do s indica e he
loca ion o dummy s a ions wi hin he ci y cen e . These
dummy s a ions a e posi ioned in a eas wi h impo an
s uc u es such as esiden ial buildings, ai po s, hospi-
als, and o he essen ial acili ies.
3 G ound mo ion simula ion me hodology
In egions lacking well-s udied local sou ce and eloc-
i y models, he s ochas ic me hodology is conside ed
e ec i e and p e e ed o simula ing g ound mo ions.
This me hod is pa icula ly ad an ageous o cap u -
ing he alea o y unce ain y associa ed wi h ea h-
quakes, he eby accoun ing o he inhe en a iabili y
in g ound mo ions. Howe e , i p ima ily ocuses on
simula ing he andom in e media e and medium- o
high- equency componen s o g ound mo ions (Be es-
ne and A kinson 1997; Da id M. Boo e 1983a, 1983b;
Mo azedian and A kinson 2005). While his is sui able
o many enginee ing applica ions, such as o s uc-
u es sensi i e o hese equency anges, he app oach
does no ully cap u e low- equency componen s.
Low- equency g ound mo ions, which a e impo an
151
J Seismol (2025) 29:147–169
Vol.: (0123456789)
o he esponse o all buildings and o he long-pe iod
s uc u es, equi e mo e complex de e minis ic phys-
ics-based simula ions o a hyb id simula ion scheme.
Despi e hese limi a ions, s ochas ic echniques a e
known o accu a ely simula e g ound mo ion ampli-
udes wi hin he equency ange o p ima y in e es o
enginee s (Boo e 2009; Ka imzadeh 2019; VahidiFa d
e al. 2017).
Fig. 1 Epicen e o ea hquakes wi hin a 20 km adius a ound he No h Tab iz Faul (NTF) (adap ed om (Fa ahani 2018) and
(Gho bani 2021))
Fig. 2 Faul line and selec ed s a ions o g ound mo ion simula ions
152
J Seismol (2025) 29:147–169
Vol:. (1234567890)
Fo he simula ion o he scena io e en s in his
s udy, he s ochas ic ini e- aul g ound mo ion sim-
ula ion echnique de eloped by (Mo azedian and
A kinson 2005) is employed using he EXSIM pla -
o m. This app oach ep esen s he aul plane using
a ec angula geome y and models i as a collec ion
o smalle sub- aul s ha ac as poin sou ces (Da id
M Boo e 1983a, 1983b). The accele a ion esponse
spec um in he equency domain is modeled o
each poin sou ce, aking in o accoun he con ibu-
ions o sou ce, pa h, and si e e ec s. Subsequen ly,
he esponses o all sub- aul s a e summed in he
ime domain o ob ain he inal esponse o he main
aul as ollows:
whe e
a( )
is he o al seismic signal a a ime , nl and
nw a e he numbe o sub- aul s along he leng h and
wid h o he main aul , espec i ely, and
Δ ij
is he
ela i e delay ime o he adia ed wa e om he ij h
sub- aul o each he obse a ion poin . Fo each sub-
aul (Da id M Boo e 1983a, 1983b), he accele a ion
spec um is compu ed as:
(1)
a
( )=
∑
nl
i=1
nw
∑
j=1
aij( + Δ ij)
,
(2)
Aij( )=CM0ij Hij
⎡⎢⎢⎣
(2𝜋 )2∕
⎡⎢⎢⎣
1+
�
cij
�
2
⎤⎥⎥⎦⎤⎥⎥⎦
e−𝜋 Rij
Q( )𝛽G
�
Rij
�
A( )e−𝜋K
Whe e
C
=
R
𝜃𝜑FV
4𝜋𝜌𝛽
2 is a cons an , wi h
R𝜃𝜑
as he adia-
ion pa e n (a e age alue o 0.55 o shea wa es, F
as ee su ace ampli ica ion (2.0), V as pa i ion on o
wo ho izon al componen s (0.71),
ρ
as he densi y,
and
β
as he shea -wa e eloci y.
M
0ij =
M
0
S
ij
∑
nw
k=1∑
nl
l=1
S
kl
is
he seismic momen ,
Sij
is he ela i e slip weigh and
c
ij ( )
is he dynamic co ne equency o he
ij h
sub-
aul whe e
ij( )=N
R( )−1∕34.9 ×106β( Δσ
M0
a e
)
1∕3
.
He e,
Δσ
is he s ess d op,
NR( )
is he cumula i e
numbe o up u ed sub- aul s a ime
, and
M
0
a e
=
M0
N
is he a e age seismic momen o sub-
aul s.
Sij
is he dis ance om he obse a ion poin ,
Q( )
is he quali y ac o ,
G(
R
ij)
is he geome ic
sp eading ac o ,
A( )
ep esen s he si e ampli ica ion
e m, and
e−πK
is a high-cu il e included o p o ide
he spec al decay a high equencies desc ibed wi h
he Kappa ac o o soils (Ande son and Hough
1984).
Rij
is a scaling ac o in oduced o conse e
he high- equency spec al le el o he sub aul s. The
e m
H
ij =(N∗(
∑
( 2∕[1+
�
0
�
2
])∕
∑
(( 2∕[1+
�
0
ij �
2
])))
1∕2
co esponds o a scaling ac o o con e sa ion o he
spec al shape a highe equencies.
To simula e he scena io e en s o each magni-
ude, he leng h and wid h o he up u ed aul a e
es ima ed using he s udy by Wells and Coppe smi h
(1994). The en i e leng h o he NTF is app oxi-
ma ely 240 km (Fa ahani 2018; Solaymani Azad
e al. 2015). Al e na i e aul planes a e conside ed
o cap u e he a iabili y o he up u e plane. Fo
he la ges e en wi h Mw = 7.7, he NTF is di ided
in o 17 al e na i e up u ed aul lines, each span-
ning a leng h o 140 km, o gene a e a ious pos-
sible scena ios. Simila ly, o he o he magni udes,
he NTF is di ided in o mo e han 13 e en s o each
case. Speci ically, o Mw = 7.4, he aul plane is
di ided in o 12 segmen s, each wi h a leng h o 84
km; o Mw = 7.1, i is di ided in o 13 segmen s wi h
a leng h o 52 km, and o Mw = 6.8 i is di ided in o
15 segmen s wi h a leng h o 32 km. The espec i e
wid hs o hese up u ed aul planes a e de e mined
o be 24 km, 20 km, 16 km, and 12 km o Mw = 7.7,
7.4, 7.1, and 6.8, espec i ely. I is no ed ha o all
scena ios wi h e en s o he same magni ude, he
epicen e is shi ed along he aul plane by an equal
dis ance.
In addi ion o he a iabili y o he up u ed aul
plane, he s udy also in es iga es he andomness
associa ed wi h he hypocen e loca ion o al e na-
i e scena io e en s. To achie e his, he ocal dep h
and epicen e loca ion a e ea ed as andom a ia-
bles. Mo adi e al. (2011) ound ha he median ocal
dep h o ea hquakes in he Tab iz egion is app oxi-
ma ely 12.1 km, wi h a s anda d de ia ion o a ound
4.9 km. In his s udy, o inco po a e he unce ain y
ela ed o he ocal dep h, a ange o 6.0 km o 18.0
km is conside ed. Addi ionally, he epicen e loca ion
is assumed o change along he leng h o he up u ed
aul plane.
153
J Seismol (2025) 29:147–169
Vol.: (0123456789)
Excluding he andom pa ame e s discussed abo e,
he o he inpu -model pa ame e s a e de e mined in
a de e minis ic manne ac oss all magni udes and
a e essen ial o comple ing he simula ions. These
pa ame e s a e sou ced om a ious s udies. The
s ike and dip angle alues, de i ed as 115° and 90°,
espec i ely, a e ob ained om he s udies by Fa a-
hani (2018) and Ghayamghamian and Rajool (2012).
The quali y ac o is aken om he s udy conduc ed
by Ja a ian Ve nos ade ani e al. (2019), while he
kappa ac o is de i ed as 0.035 based on he esea ch
conduc ed by Ami anlou e al. (2016). The du a-
ion model ollows he wo k o He mann (1985) as
applied in Ho eidae e al. (2021) o he simula ion o
scena io ea hquakes in he Tab iz egion. This s udy
uses si e ampli ica ions om Boo e and Joyne (1997)
o a gene ic soil class wi h a Vs30 alue o 310 m/s.
This alue is selec ed because he dummy s a ions
a e mainly loca ed in a eas wi h Eu ocode soil Class
C (Code P ice 2005), wi h Vs30 alues anging om
300 o 360 m/s, as indica ed by GIS and DEM-based
seismic si e condi ion s udies by Ka imzadeh e al.
(2017, 2019). Yaghmaei-Sabegh and Hassani (2020)
u he co obo a e his by epo ing an a e age Vs30
alue o app oxima ely 300 m/s o he egion, e lec -
ing minimal soil a iabili y based on analyses o seis-
mic e en s in I an. Al hough soil a iabili y is impo -
an in imp o ing g ound mo ion simula ions, his
s udy adop s a cons an Vs30 alue o ocus on unce -
ain ies ela ed o he aul up u e plane and hypo-
cen e loca ion, which limi s he esul s o a gene ic
soil classi ica ion. The s ess d op is calcula ed using
he empi ical equa ion gi en by Mohammadioun and
Se a (2001). Geome ical sp eading, pulsing pe -
cen , up u e eloci y, densi y, and window ype a e
assumed o be he same as in he esea ch conduc ed
by Ami anlou e al. (2016) and Ho eidae e al. (2021).
A summa y o all his in o ma ion can be ound in
Table1.
4 Simula ed g ound mo ion da ase
A o al o 206,382 simula ed g ound mo ion ime
se ies ha e been gene a ed o Tab iz Ci y a he
selec ed s a ions. The da ase encompasses ea h-
quake scena ios wi h Mw o 6.8, 7.1, 7.4, and 7.7,
and Joyne -Boo e dis ances (RJB) anging om 5.4
km o 87 km wi h ocal dep hs (Fd) be ween 6 and
18 km. Gi en ha soil class C p edomina es in he
egion, we used a ep esen a i e Vs30 alue o 310
m/s unde a s ike-slip aul mechanism. The seis-
mological pa ame e s o he da ase a e illus a ed in
Fig.3. I is no ewo hy ha a highe numbe o simu-
la ions ha e been gene a ed o la ge Mw alues,
sho e RJB dis ances, and shallowe ocal dep hs o
add ess he sca ci y o s ong nea - aul eco dings.
Figu e4 p esen s sample simula ed g ound mo ions
o scena io e en s wi h Mw o 7.7, 7.4, 7.1, and 6.8,
a a RJB dis ance o 12 km, and a Fd o 10 km along
wi h hei espec i e peak g ound accele a ion (PGA)
alues.
5 Assessmen o simula ions
This sec ion ocuses on alida ing he simula ions
h ough compa ison wi h GMMs, analyzing a enu-
a ion o he simula ed da a, and de eloping seismic
haza d and MMI maps.
Table 1 De e minis ic inpu pa ame e s o g ound mo ion
simula ion o scena io e en s wi h di e en magni ude alues
Pa ame e s Value
Mw7.7 7.4 7.1 6.8
Leng h (km) 140 84 52 32
Wid h (km) 24 20 16 12
Sub- aul leng h (km) 4 4 4 4
Sub- aul wid h (km) 4 4 4 4
S ess D op (ba s) 110 95 80 65
Faul Mechanism S ike-slip
S ike (o) 115
Dip (o) 90
Shea Wa e Veloci y (km/s) 3.3
Rup u e Veloci y/Shea Wa e Veloci y 0.8
Densi y (g/cm3) 2.8
Pulsing Pe cen 35
Quali y Fac o
103 0.88
Geome ical Sp eading
R−1R≤85
R−0.5R>85
Du a ion model T = T0 + 0.05
(T0: Sou ce du a-
ion)
Kappa Value 0.035
Window Type Sa agoni-Ha
Si e Ampli ica ions Gene ic soil
Sampling Time (s) 0.005
154
J Seismol (2025) 29:147–169
Vol:. (1234567890)
5.1 Valida ion o simula ions
I is impo an o alida e he simula ed mo ions
be o e hey can be used o enginee ing applica ions.
Valida ion me hods a y depending on he a ailabili y
o eco ded mo ions. In cases whe e eco ded mo ions
a e absen , a commonly employed app oach in ol es
using empi ical GMMs o alida ion pu poses. The
Fig. 3 His og ams o seismological pa ame e s o he simula ed da ase
Fig. 4 Samples o simula ed ime se ies o he s udy a ea
155
J Seismol (2025) 29:147–169
Vol.: (0123456789)
li e a u e o e s nume ous pa ame ic and non-pa a-
me ic GMMs. Howe e , selec ing app op ia e GMMs
o compa ison is c ucial, especially hose ha exhibi
simila seismological and ec onic cha ac e is ics
o he egion unde s udy. In his s udy, ou dis inc
GMMs a e used o he alida ion p ocess, namely
BA08 (Boo e and A kinson 2008), AC10 (Akka
and Cagnan 2010), ASB14 (Akka e al. 2014), and
KAAH15 (Kale e al. 2015). These ou selec ed
GMMs ul ill di e en pu poses. The BA08 model is
ecognized as a global GMM due o i s de elopmen ,
inco po a ing eco ded da a om di e se global loca-
ions. In con as , he AC10 model se es as a local
model, speci ically ailo ed o he seismic cha ac e is-
ics o Tü kiye. The ASB14 model is a egional one
de eloped o cap u e he seismic a ibu es o he Mid-
dle Eas e n and Eu opean egions, encompassing mul-
iple coun ies. Las ly, he KAAH15 model alls unde
he ca ego y o a local GMM, cus omized o accoun
o he seismic beha io obse ed in I an and Tü -
kiye. The alida ion p ima ily ocuses on i e essen ial
in ensi y measu es (IMs): PGA, Peak G ound Veloci y
(PGV), and Pseudo-Spec al Accele a ion (PSA) wi h
a 5% damping a io o pe iods o T = 0.2 s, 0.5 s, and
1.0 s. These pa ame e s se e as ep esen a i e indica-
o s o e alua ing he accu acy and eliabili y o he
simula ed da ase .
Figu e 5 illus a es he dis ance a enua ion o
simula ed g ound mo ion IMs (i.e., PGA, PGV, and
PSA a T = 0.2 s, 0.5 s, and 1.0 s) in compa ison o
es ima es om he GMMs. The esul s indica e ha
he simula ed IMs gene ally all wi hin ± 2 s and-
a d de ia ions (s d) o he GMM es ima es. Fo dis-
ances less han 20 km, he simula ed in ensi y le els
closely align wi h all GMMs. Howe e , he a enua-
ion pa e ns o simula ed IMs a e no iceably as e
a dis ances exceeding 20 km, pa icula ly o longe -
pe iod PSA. These disc epancies can be a ibu ed o
a ious pa ame e se ings in he ini e aul s ochas ic
simula ion me hod. Key ac o s, such as aul geome-
y (leng h, wid h, and dep h), s ess d op, and up u e
eloci y, in luence how seismic ene gy is eleased
and p opaga ed. The choice o geome ical sp eading
model signi ican ly a ec s how IMs decay wi h dis-
ance, while he quali y ac o and kappa alue ep e-
sen ene gy dissipa ion in he medium and si e e ec s,
espec i ely. O e all, hese pa ame e se ings con-
ibu e o he obse ed di e ences be ween simula ed
IMs and hose om empi ical GMMs, unde sco ing
he impo ance o conduc ing egion-speci ic s ud-
ies o e ine hese pa ame e s. The p esen esea ch
a emp s o assess no only he a enua ion o he ep-
esen a i e IMs bu also o de e mine he an icipa ed
seismic haza d le el wi hin he s udy a ea.
As a u he s ep in he alida ion p ocess, he co -
ela ions be ween no malized esidual (
ϵ
) o PSA
alues o a ious ib a ion pe iods a e in es iga ed.
Recen ly, Al indal and Askan (2023) alida ed a sim-
ula ed g ound mo ion da ase gene a ed o he en i e
Tü kiye wi h his app oach yielding a obus assess-
men . Gi en a PSA om an ea hquake scena io wi h
magni ude Mw a a eco ded dis ance RJB,
ϵ
alue can
be calcula ed as (Bake 2011):
whe e
μlnPSA(Mw,RJB,T)
and
σlnPSA(T)
a e he mean
alue and s anda d de ia ion o lnPSA(T) es ima ed
om GMM, espec i ely. lnPSA(T) is he na u al
loga i hm o obse ed PSA alues, e e ed o he ein
as he simula ion. In his s udy, he
ϵ
alues a e calcu-
la ed o all he simula ed mo ions in he da ase o
a pe iod ange o 0.05 s o 2 s, ega ding each o he
selec ed GMMs. In seismic demand assessmen o
mul i-deg ee-o - eedom (MDOF) sys ems ha a e
subjec ed o a ious pe iods o exci a ion, he co ela-
ion coe icien o
ϵ
ac oss he pe iod ange (
ρϵ
) is a
c ucial indica o (Bu ks and Bake 2014). The e o e,
ρϵ
o PSA, ob ained using Pea son p oduc -momen
co ela ion coe icien , ega ding each selec ed GMM
is illus a ed in Fig. 6 o e he conside ed pe iod
ange. These con ou s indica e ha he co ela ions
calcula ed based on BA08 exhibi he highes in e -
pe iod co ela ion, whe eas hose de i ed om AC10
and KAAH15 yield he lowes co ela ions. Howe e ,
ρϵ
alues ega ding ASB14 exhibi high in e -pe iod
co ela ions owa ds longe pe iods. None heless, a
consis en pa e n is obse ed ac oss all GMMs o
bo h in e -pe iod co ela ions and a enua ions o
conside ed in ensi y measu es, a i ming he alidi y
o he simula ed da ase .
5.2 Seismic haza d and MMI maps
To p o ide a comp ehensi e unde s anding o g ound
mo ion in ensi y le els o di e en scena io e en s,
seismic haza d maps a e de eloped o he egion.
(3)
ϵ(T) = lnPSA(T)−μ
lnPSA
(M
w
,R
JB
,T)
σ
lnPSA
(T)
162
J Seismol (2025) 29:147–169
Vol:. (1234567890)
espec i ely. The be ween-e en esiduals luc ua e
wi hin he ange o –0.4 o 0.4, whe eas he wi hin-
e en esiduals exhibi a wide span be ween –1.0 and
1.0, consis en wi h indings om p e ious s udies
(Akka e al. 2014; Kale e al. 2015; Ka imzadeh
e al. 2023a; Mohammadi e al. 2023). I is impo an
o no e ha he model’s p edic ions a e mos eli-
able wi hin he aining da a ange (6.8 ≤ Mw ≤ 7.7,
Fig. 14 Dis ibu ion o
be ween-e en (τ) and
wi hin-e en (σ) unce ain-
ies, along wi h o al unce -
ain ies (ϕ), o ln(PGA),
ln(PGV), and ln(PSA) a
pe iods anging om 0.05
s o 2 s
Fig. 15 Dis ibu ion o
he be ween-e en esidual
(η) wi h espec o Mw o
ln(PGA), ln(PGV), and
ln(PSA) a pe iods o 0.2 s,
0.5 s, 1.0 s, and 2.0 s
163
J Seismol (2025) 29:147–169
Vol.: (0123456789)
4 km ≤ RJB ≤ 54 km) and can be less accu a e when
applied ou side o ha ange.
The ed lines in Figs.15 and 16 ep esen he mean
o esiduals wi h espec o he independen a iables.
The absence o disce nible ends in he mean esidu-
als o bo h be ween-e en and wi hin-e en ca ego-
ies ac oss all IMs indica es he unbiased na u e o
model e o s. The ex emely na ow con idence in e -
als, ende ing hem in isible on he plo , add u -
he weigh o his obse a ion. The e o e, he model
shows no sys ema ic bias owa ds Mw and RJB.
7.2 Valida ion o he ANN-based GMM
The model adequacy is a ained by e alua ing he
a enua ion pa e n o IMs ac oss a ange o Mw
and RJB alues wi h a Vs30 o 310 m/s, a s ike-slip
mechanism, and a mean Fd o 10 km. Figu e17 illus-
a es he a ia ion o selec ed IMs wi h espec o Mw
o di e en RJB alues o 5.0, 20.0, and 50.0 km. IMs
acqui ed om simula ed ea hquake e en s a e ep e-
sen ed by illed g ay do s. Addi ionally, a compa ison
wi h KAAH15, a pa ame ic GMM ailo ed o I an,
is p esen ed o u he analysis. Resul s indica e
ha highe Mw and lowe RJB alues co ela e wi h
ele a ed le els o all selec ed IMs. Simila ly, Fig.18
demons a es he a ia ion o hese IMs wi h RJB, o
di e en Mw alues o 7.0, 7.3, and 7.6, alongside a
compa ison o he KAAH15 GMM. I is obse ed
ha as RJB inc eases, he e is a consis en dec ease in
PGA, PGV, and PSA ac oss all pe iods, a end e ec-
i ely cap u ed by he p oposed GMM, highligh ing
Fig. 16 Dis ibu ion o
he wi hin-e en esidual
(ϵ) wi h espec o RJB o
ln(PGA), ln(PGV), and
ln(PSA) a pe iods o 0.2 s,
0.5 s, 1.0 s, and 2.0 s
164
J Seismol (2025) 29:147–169
Vol:. (1234567890)
Fig. 17 Va ia ion o PGA, PGV, and PSA a pe iods o 0.2, 0.5, 1.0, and 2.0 s wi h espec o Mw o a s ike-slip aul mechanism,
Vs30 = 310 m/s, and RJB alues o 5.0, 20.0, and 50.0 km. The p oposed model is shown wi hin wo s anda d de ia ions
165
J Seismol (2025) 29:147–169
Vol.: (0123456789)
Fig. 18 Va ia ion o PGA, PGV, and PSA a pe iods o 0.2, 0.5, 1, and 2 s wi h espec o RJB o a s ike-slip aul mechanism,
Vs30 = 310 m/s, and Mw alues o 7.0, 7.3, and 7.6 km. The p oposed model is shown wi hin wo s anda d de ia ions
166
J Seismol (2025) 29:147–169
Vol:. (1234567890)
i s abili y o e lec dis ance-dependen a enua-
ion. Mo eo e , he ela ionship be ween highe Mw
and ampli ied IMs emains consis en wi h p io
obse a ions, as he magni ude inc ease co ela es
wi h a g ea e ene gy elease du ing seismic e en s
(Ka imzadeh e al. 2023a). The analysis e eals a
s ong alignmen be ween he p oposed ANN-based
GMM using simula ed da a and he empi ical GMM
de i ed om eal da a, no ably in PGA and PSA
a he pe iod o 0.2 s. None heless, disc epancies
become mo e p onounced o o he IMs, pa icula ly
no iceable in PSA a longe pe iods, whe e he dispa -
i y inc eases. Some o hese disc epancies migh be
a ibu ed o he selec ion o he model pa ame e s in
he simula ions as well as he po en ial limi a ions o
he s ochas ic ini e- aul app oach in accu a ely cap-
u ing he longe pe iod con en o he eco ds and
basin e ec s.
8 Conclusions
This s udy employs s ochas ic ini e- aul g ound
mo ion simula ion based on he dynamic co ne e-
quency o simula e al e na i e scena io e en s a di e -
en magni ude alues wi hin he ci y cen e o Tab iz,
I an. The e ec s o di e se hypocen e loca ions and
up u ing aul planes o speci ic magni udes a e e al-
ua ed s a is ically in de ail. The analyses ocused on he
ci y cen e , allowing o he assessmen o a ia ions in
g ound mo ion in ensi y esul ing om di e en hypo-
cen e and aul plane con igu a ions.
The esul s a e alida ed agains empi ical g ound-
mo ion models (GMMs), namely BA08 (Boo e and
A kinson 2008), AC10 (Akka and Cagnan 2010),
ASB14 (Akka e al. 2014), and KAAH15 (Kale e al.
2015). The indings e eal he alidi y o he simula-
ions om a seismological poin o iew, exhibi ing
conside able a iabili y in g ound mo ion in ensi y
measu es (IMs) ac oss each si e o in e es . This a i-
abili y is a ibu ed o unce ain ies in he ea hquake
hypocen e and he aul plane up u e. Mo eo e ,
he e alua ion o he in e -pe iod co ela ion o he
esidual pseudo-spec al accele a ion (PSA) se es
as an addi ional s ep in he alida ion p ocess. The
highes alues o all g ound mo ion IMs a e consis -
en ly obse ed in he immedia e icini y o he No h
Tab iz Faul (NTF) line, unde sco ing he signi i-
cance o his a ea in e ms o po en ial haza ds.
The indings sugges ha highe ea hquake mag-
ni udes esul in highe alues o seismic IMs such
as peak g ound accele a ion (PGA), peak g ound
eloci y (PGV), and PSA ac oss di e en pe iods.
Speci ically, compa ing he momen magni ude (Mw)
o 7.7 wi h ha o 6.8, 7.1, and 7.4 e eals a signi i-
can educ ion o app oxima ely 65%, 45%, and 15%,
espec i ely, based on he mean a e age o he IMs.
Fu he mo e, he conside able s anda d de ia ions
conce ning he mean alues ac oss a ious Mw and
IMs indica e a b oad spec um o po en ial g ound-
shaking le els. The a ying seismic in ensi y expe-
ienced du ing ea hquakes o simila magni udes
wi hin he egion unde sco es he signi ican haza d
aced by he ci y cen e , especially o he highes
Mw o 7.7, s essing he impo ance o p epa edness.
Es ima ions o Modi ied Me calli in ensi y (MMI)
alues ac oss a ious scena ios u he highligh he
inc eased ulne abili y o Tab iz ci y cen e o signi i-
can po en ial haza ds.
A local g ound mo ion model (GMM) is de el-
oped by aining a i icial neu al ne wo ks (ANN)
on simula ed g ound mo ions. The de eloped ANN-
based GMM inco po a es momen magni ude (Mw),
Joyne -Boo e dis ance (RJB), and ocal dep h (Fd) as
inpu pa ame e s, enabling he es ima ion o ln(PGA),
ln(PGV), and ln(PSA) a a ious pe iods. Th ough
esidual analysis, he unce ain y o he GMM is quan-
i ied in e ms o be ween-e en and wi hin-e en a i-
a ions. Compa ison wi h an empi ical GMM ailo ed
o I an e eals he obus pe o mance o he p oposed
GMM in accu a ely cap u ing obse ed pa e ns in
g ound mo ions despi e some disc epancies o spec al
o dina es a longe pe iods. This migh be a ibu ed
o he po en ial limi a ions o he s ochas ic ini e- aul
app oach in accu a ely cap u ing he longe pe iod
con en o he eco ds and basin e ec s. Mo eo e ,
an online pla o m is o e ed o ensu e a use - iendly
applica ion o he de eloped GMM o end-use s.
The analyses in his s udy a e pe o med wi h
uni o m soil condi ions. Fu u e s udies should aim
o add ess he unce ain y s emming om inpu
pa ame e s, pa icula ly si e pa ame e s along wi h
he simula ion me hodology u he o enhance ou
unde s anding o seismic haza ds in he egion. Addi-
ionally, u u e s udies could explo e he in eg a ion
o hese indings wi h ulne abili y and exposu e
da a o de elop a comp ehensi e isk assessmen
amewo k.
167
J Seismol (2025) 29:147–169
Vol.: (0123456789)
Acknowledgemen s This wo k was pa ly inanced by FCT
/ MCTES h ough na ional unds (PIDDAC) unde he R&D
Uni Ins i u e o Sus ainabili y and Inno a ion in S uc u al
Enginee ing (ISISE), unde e e ence UIDB / 04029/2020 (doi.
o g/h ps:// doi. o g/ 10. 54499/ UIDB/ 04029/ 2020), and unde
he Associa e Labo a o y Ad anced P oduc ion and In elligen
Sys ems (ARISE) unde e e ence LA/P/0112/2020. This s udy
has been pa ly unded by he STAND4HERITAGE p ojec
ha has ecei ed unding om he Eu opean Resea ch Coun-
cil (ERC) unde he Eu opean Union’s Ho izon 2020 esea ch
and inno a ion p og am (G an Ag eemen No. 833123), as an
Ad anced G an . This wo k is pa ly inanced by na ional unds
h ough FCT (Founda ion o Science and Technology), unde
g an ag eemen UI/BD/153379/2022 a ibu ed o he second
au ho .
Au ho con ibu ion C. Temiz: Fo mal analysis, Da a cu a-
ion, In es iga ion, Resou ces, Visualisa ion, W i ing—o iginal
d a , W i ing— e iew & edi ing. S. M. S. Hussaini: Fo mal
analysis, In es iga ion, Visualisa ion, W i ing—o iginal d a ,
W i ing— e iew & edi ing. S. Ka imzadeh: Concep ualisa-
ion, Da a cu a ion, Fo mal analysis, In es iga ion, Me hodol-
ogy, Resou ces, Supe ision, Valida ion, W i ing—o iginal
d a , W i ing— e iew & edi ing. A. Askan: Resou ces, W i -
ing— e iew & edi ing. P. B. Lou enço: Funding acquisi ion,
Resou ces, Supe ision, W i ing— e iew & edi ing.
Funding Open access unding p o ided by FCT|FCCN
(b-on).
Da a a ailabili y This s udy uses he S eamli package in
Py hon o cons uc a use - iendly g aphical in e ace ool,
acili a ing con enien access o he ANN-based g ound mo ion
model (GMM). The code can be ound a h ps://gi hub.com/S-
M-S–H/Tab iz-GMM-ANN, while he in e ace ool i sel is
accessible ia h ps:// ab iz-gmm-ann.s eamli .app/. Use s a e
p omp ed o inpu he pa ame e s o a scena io ea hquake,
such as momen magni ude (Mw), Joyne -Boo e dis ance
(RJB), and ocal dep h (Fd). The so wa e hen gene a es ou -
comes in e ms o in ensi y measu es (IMs), encompassing
peak g ound accele a ion (PGA), peak g ound eloci y (PGV),
and pseudo spec al accele a ion (PSA). Finally, all da a sup-
po ing his pape will be made a ailable upon eques o he
co esponding au ho .
Decla a ions
Compe ing In e es s The au ho s decla e no compe ing
in e es s.
Open Access This a icle is licensed unde a C ea i e Com-
mons A ibu ion 4.0 In e na ional License, which pe mi s
use, sha ing, adap a ion, dis ibu ion and ep oduc ion in any
medium o o ma , as long as you gi e app op ia e c edi o he
o iginal au ho (s) and he sou ce, p o ide a link o he C ea-
i e Commons licence, and indica e i changes we e made. The
images o o he hi d pa y ma e ial in his a icle a e included
in he a icle’s C ea i e Commons licence, unless indica ed
o he wise in a c edi line o he ma e ial. I ma e ial is no
included in he a icle’s C ea i e Commons licence and you
in ended use is no pe mi ed by s a u o y egula ion o exceeds
he pe mi ed use, you will need o ob ain pe mission di ec ly
om he copy igh holde . To iew a copy o his licence, isi
h p://c ea i ecommons.o g/licenses/by/4.0/.
Re e ences
Ab ahamson NA, Youngs RR (1992) A s able algo i hm o
eg ession analyses using he andom e ec s model. Bull
Seismol Soc Am 82(1):505–510
Akka S, Cagnan Z (2010) A local g ound-mo ion p edic i e
model o Tu key, and I s compa ison wi h o he egional
and global g ound-mo ion models. Bull Seismol Soc Am
100(6):2978–2995
Akka S, Sandıkkaya MA, Bomme JJ (2014) Empi ical
g ound-mo ion models o poin - and ex ended-sou ce
c us al ea hquake scena ios in Eu ope and he Middle
Eas . Bull Ea hq Eng 12(1):359–387 (Kluwe Academic
Publishe s)
Al indal A, Askan A (2023) Cons uc ion and alida ion o a
simula ed g ound mo ion da ase o Tu key. J Seismol
2023:1–19 (Sp inge )
Ami anlou H, Pou ke mani M, Dabi i R e al (2016) The s o-
chas ic ini e- aul modeling based on a dynamic co ne
equency simula ing o s ong g ound mo ion o ea h-
quake scena io o no h Tab iz aul . Open J Ea hq Res
05(02):114–121
Ande son JG, Hough SE (1984) A model o he shape o he
ou ie ampli ude spec um o accele a ion a high e-
quencies. Bull Seismol Soc Am 74(5):1969–1993
A o a S, Joshi A, Kuma i P e al (2020) S ong g ound mo ion
simula ion echniques—a e iew in wo ld con ex . A ab J
Geosci 13(14):1–12 (Sp inge )
A slanKelam A, Ka imzadeh S, Youse iba il K e al (2022)
An e alua ion o seismic haza d and po en ial damage in
Gazian ep, Tu key using si e speci ic models o sou ces,
eloci y s uc u e and building s ock. Soil Dyn Ea hq
Eng 154:107129
Askan A, Al indal A, Aydin MF e al (2024) Assessmen o
u ban seismic esilience o a own in Eas e n Tu kiye:
Tu koglu, Kah amanma as be o e and a e M7.8 Kah a-
manma as ea hquake, Ea hquake spec a.h ps:// doi. o g/
10. 1177/ 87552 93024 12747 15
A kinson GM, Goda K, Assa ou ians K (2011) Compa i-
son o nonlinea s uc u al esponses o accele og ams
simula ed om he s ochas ic ini e- aul app oach e sus
he hyb id b oadband app oach. Bull Seismol Soc Am
101(6):2967–2980 (GeoScienceWo ld)
Aza a za M, Ghazi a d A (2016) U ban geology o Tab iz
Ci y: en i onmen al and geological cons ain s. Ad En i-
on Res 5(2):95–108 (Techno‑P ess)
Bake JW (2011) Condi ional mean spec um: ool o g ound-
mo ion selec ion. J S uc Eng 137(3):322–331
Be be ian M, and A shadi S (1976) On he e ide ne o he
younges ac i i y o he No h Tab iz Faul and he seis-
mici y o Tab iz Ci y, Geological Su ey o I an, Teh an.
I an Rep 39:397–418
168
J Seismol (2025) 29:147–169
Vol:. (1234567890)
Be be ian M, Yea s RS (1999) Pa e ns o His o ical Ea h-
quake Rup u e in he I anian Pla eau. Bulle in o he Seis-
mological Socie y o Ame ica89(1):120–139.h ps:// doi.
o g/ 10. 1785/ BSSA0 89001 0120
Be esne IA, A kinson GM (1997) Modeling ini e- aul
adia ion om he ω n spec um. Bull Seismol Soc Am
87(1):67–84 (The Seismological Socie y o Ame ica)
Bilal M, Askan A (2014) Rela ionships be ween el in ensi y
and eco ded g ound-mo ion pa ame e s o Tu key. Bull
Seismol Soc Am 104(1):484–496
Boo e DM (1983a) S ochas ic simula ion o high- equency
g ound mo ions based on seismological models o he
adia ed spec a. Bull Seismol Soc Am 73(6A):1865–1894
Boo e DM (1983b) S ong-mo ion seismology. Re Geophys
21(6):1308–1318
Boo e DM (2009) Compa ing s ochas ic poin -sou ce and
ini e-sou ce g ound-mo ion simula ions: SMSIM and
EXSIM. Bull Seismol Soc Am 99(6):3202–3216
Boo e DM, A kinson GM (2008) G ound-mo ion p edic ion
equa ions o he a e age ho izon al componen o PGA,
PGV, and 5%-damped PSA a spec al pe iods be ween
0.01 s and 10.0 s. Ea hq Spec a 24(1):99–138 (Ea h‑
quake Enginee ing Resea ch Ins i u e)
Boo e DM, Joyne WB (1997) Si e Ampli ica ions o Gene ic
Rock Si es. Bulle in o he Seismological Socie y o
Ame ica 87(2):327–341. h ps:// doi. o g/ 10. 1785/ BSSA0
87002 0327
Bu ks LS, Bake JW (2014) Valida ion o g ound-mo ion simu-
la ions h ough simple p oxies o he esponse o engi-
nee ed sys ems. Bull Seismol Soc Am 104(4):1930–1946
(GeoScienceWo ld)
Can G, Askan A, Ka imzadeh S (2021) An assessmen o he 3
Feb ua y 2002 Cay (Tu key) ea hquake (Mw=6.6): mod-
eling o g ound mo ions and el in ensi y dis ibu ion.
Soil Dyn Ea hq Eng 150:106832
Dhanya J, Raghukan h STG (2018) G ound mo ion p edic ion
model using A i icial Neu al Ne wo k. Pu e Appl Geo-
phys 175(3):1035–1064 (Bi khause Ve lag AG)
Code P ice (2005) Eu ocode 8: Design o s uc u es o ea h-
quake esis ance-pa 1: gene al ules, seismic ac ionsand
ules o buildings. Eu opean Commi ee o S anda diza-
ion, B ussels, p 10
Eslami M, Taghizadeh-Fa ahmand F (2020) E alua ion o he
de e minis ic seismic haza d by using uzzy in e ence sys-
em, case s udy: Tab iz Ci y, I an. I an J Geophys (IJG)
14(4):19–28
Fa ahani JV (2018) S a is ical p ope ies o a e shocks o
Aha -Va zeghan win ea hquakes on 11 Augus 2012,
NW I an, and in es iga ion o seismici y o no h Tab iz
aul . In J Geosci 09(02):106–118
F ankel A (1993) Th ee-dimensional simula ions o g ound
mo ions in he San Be na dino Valley, Cali o nia, o
hypo he ical ea hquakes on he San And eas aul . Bull
Seismol Soc Am 83(4):1020–1041
Ghasemi M, Ka imzadeh S, Ma suoka M e al (2021) Wha
would happen i he M 7.3 (1721) and M 7.4 (1780) his-
o ical ea hquakes o Tab iz Ci y (NW I an) occu ed
again in 2021? ISPRS In J Geo-In 10(10):657
Ghayamghamian MR, Rajool A (2012) Long-pe iod g ound
mo ion simula ion o NTF aul nea -sou ce ene gy
eleased. P oc Fi een h Wo ld Con Ea hq Eng, Lisbon,
Po ugal 2012:23–25
Gho bani M (2021) Faul s and ec onic phases o I an, pp.
81–149. h ps:// doi. o g/ 10. 1007/ 978-3- 030- 71109-2_3
Hancock J, Wa son-Lamp ey J, Ab ahamson NA e al (2006)
An imp o ed me hod o ma ching esponse spec a o
eco ded ea hquake g ound mo ion using wa ele s. J
Ea hq Eng 10(sup001):67–89 (Wo ld Scien i ic)
Ha i i-A debili MA, Rezaeian S (2024) U iliza ion o s ochas-
ic g ound mo ion simula ions o scena io-based pe o -
mance assessmen o geo-s uc u es. Reliab Eng Sys Sa
251:110375 (Else ie )
He mann RB (1985) An ex ension o andom ib a ion heo y
es ima es o s ong g ound mo ion o la ge dis ances. Bull
Seismol Soc Am 75(5):1447–1453
Hessami K, Pan os i D, Tabassi H e al (2003) Paleoea h-
quakes and slip a es o he No hTab iz Faul , NW I an:
p elimina y esul s. Ann Geophys 45(5). h ps:// doi. o g/
10. 4401/ ag- 3461
Hessami K, Pan os i D, Tabassi H e al (2009) Paleoea h-
quakes and slip a es o he No h Tab iz Faul , NW I an:
p elimina y esul s. Ann Geophys 46(5). h ps:// doi. o g/
10. 4401/ ag- 3461
Ho eidae N, Fa hi A, Ka imzadeh S (2021) Seismic damage
assessmen o a his o ic mason y building unde simu-
la ed scena io ea hquakes: a case s udy o A ge-Tab iz.
Soil Dyn Ea hq Eng 147:106732 (Else ie )
Ja a ianVe nos ade ani SL, Heida i R, Mi zaei N e al (2019)
Coda wa e a enua ion in he no hwes e n I an, using
sho ime Fou ie ans o m. J Seismol 23(5):1085–1095
(Sp inge Ne he lands)
Kale Ö, Akka S, Ansa i A e al (2015) A g ound-mo ion p edic-
i e model o I an and Tu key o ho izon al PGA, PGV,
and 5% damped esponse spec um: in es iga ion o possible
egional e ec s. Bull Seismol Soc Am 105(2A):963–980
Kamae K, I iku a K, Pi a ka A (1998) A echnique o simula -
ing s ong g ound mo ion using hyb id G een’s unc ion.
Bull Seismol Soc Am 88(2):357–367
Ka akhanian AS, T i ono VG, Philip H e al (2004) Ac i e
aul ing and na u al haza ds in A menia, eas e n Tu key
and no hwes e n I an. Tec onophysics 380(3–4):189–219
Ka imzadeh S, Askan A (2018) Modeling o a his o ical
ea hquake in E zincan, Tu key (Ms~7.8, in 1939) using
egional seismological in o ma ion ob ained om a ecen
e en . Ac a Geophys 66(3):293–304
Ka imzadeh S, Miyajima M, Hassanzadeh R e al (2014) A
GIS-based seismic haza d, building ulne abili y and
human loss assessmen o he ea hquake scena io in
Tab iz. Soil Dyn Ea hq Eng 66:263–280
Ka imzadeh S, Feizizadeh B, Ma suoka M (2017) F om a
GIS-based hyb id si e condi ion map o an ea hquake
damage assessmen in I an: me hods and ends. In J
Disas e Risk Reduc ion 22:23–36
Ka imzadeh S, Feizizadeh B, Ma suoka M (2019) DEM-
based Vs30 map and e ain su ace classi ica ion in
na ionwide scale—a case s udy in I an. ISPRS In J Geo
In 8(12):537
Ka imzadeh S, Mohammadi A, Hussaini SMS e al (2023)
ANN-based g ound mo ion model o Tu key using
s ochas ic simula ion o ea hquakes. Geophys J In
236(1):413–429 (Ox o d Academic)
169
J Seismol (2025) 29:147–169
Vol.: (0123456789)
Ka imzadeh S, Funa i MF, Szabó S e al (2024) S ochas ic
simula ion o ea hquake g ound mo ions o he seismic
assessmen o monumen al mason y s uc u es: sou ce-
based s si e-based app oaches. Ea hq Eng S uc Dyn
53(1):303–330 (John Wiley & Sons)
Ka imzadeh S, Mohammadi A, Salahuddin U e al (2023)
Backbone g ound mo ion model h ough simula ed
eco ds and XGBoos machine lea ning algo i hm: an
applica ion o he Azo es pla eau (Po ugal). Ea hq
Eng S uc Dyn. John Wiley & Sons, L d. h ps:// doi.
o g/ 10. 1002/ EQE. 4040
Ka imzadeh S (2019) Seismological and enginee ing demand
mis i s o e alua ing simula ed g ound mo ion eco ds.
Appl Sci (Swi ze land) 9(21). MDPI AG. h ps:// doi.
o g/ 10. 3390/ app92 14497
Khei izadeh A ouq M, Esmaeilpou M, Sa a H (2020) Vul-
ne abili y assessmen o ci ies o ea hquake based on
he ca as ophe heo y: a case s udy o Tab iz ci y, I an.
En i on Ea h Sci 79(14). Sp inge . h ps:// doi. o g/ 10.
1007/ s12665- 020- 09103-2
Khos a ikia F, Clay on P (2021) Machine lea ning in
g ound mo ion p edic ion. Compu Geosci 148:104700
(Pe gamon)
Ma qua d DW (1963) An algo i hm o leas -squa es es i-
ma ion o nonlinea pa ame e s. J Soc Ind Appl Ma h
11(2):431–441
Mohammadi A, Ka imzadeh S, Valizadeh Kam an K e al
(2020a) Ex ac ion o land in o ma ion, u u e landscape
changes and seismic haza d assessmen : a case s udy o
Tab iz, I an. Senso s 20(24):7010
Mohammadi A, Ka imzadeh S, Kam an KV e al (2020)
Ex ac ion o land in o ma ion, u u e landscape changes
and seismic haza d assessmen : a case s udy o ab iz,
i an. Senso s (Swi ze land) 20(24):1–27 (MDPI AG)
Mohammadi A, Ka imzadeh S, Ami S e al (2023) The
po en ial o egion-speci ic machine-lea ning-based
g ound mo ion models: applica ion o Tu key. Soil Dyn
Ea hq Eng 172(Janua y):108008 (Else ie )
Mohammadioun B, Se a L (2001) S ess d op, slip ype,
ea hquake magni ude, and seismic haza d. Bull Seismol
Soc Am 91(4):694–707 (GeoScienceWo ld)
Mo adi AS, Ha z eld D, Ta a M (2011) Mic oseismici y and
seismo ec onics o he No h Tab iz aul (I an). Tec-
onophysics 506(1–4):22–30
Mo azedian D, A kinson GM (2005) S ochas ic ini e- aul
modeling based on a dynamic co ne equency. Bull Seis-
mol Soc Am 95(3):995–1010
Ozsa ac V, Ka imzadeh S, Askan A e al (2021) Seismic
demands o ba e and base-isola ed s eel ames o
eal agains simula ed eco ds o a pas ea hquake.
18(9):1266–1281.h ps:// doi. o g/ 10. 1080/ 15732 479. 2021.
18952 27 Taylo & F ancis
Ped egosa F, Va oquaux G, G am o A e al (2011) Sciki -
lea n: machine lea ning in Py hon. J Mach Lea n Res
12(85):2825–2830
Rezaeian S, S ewa JP, Luco N e al (2024) Findings om a
decade o g ound mo ion simula ion alida ion esea ch
and a pa h o wa d. Ea hq Spec a 40(1):346–378 (SAGE
Publica ions Inc.)
Rezaeian S, Sun X (2014) S ochas ic g ound mo ion simula-
ion. Encyclopedia o Ea hquake Enginee ing. Be lin,
Heidelbe g: Sp inge Be lin Heidelbe g: 1–15. h ps:// doi.
o g/ 10. 1007/ 978-3- 642- 36197-5_ 239-1
Sme zini C, Amendola C, Paolucci R e al (2024) Enginee -
ing alida ion o BB-SPEEDse , a da a se o nea -sou ce
physics-based simula ed accele og ams. Ea hq Spec a
40(1):420–445 (SAGE Publica ions Inc.)
Solaymani Azad S, Philip H, Dominguez S e al (2015) Paleo-
seismological and mo phological e idence o slip a e
a ia ions along he no h ab iz aul (NW I an). Tec ono-
physics 640:20–38 (Else ie )
Tanı can G, Yelkenci-Necmioğlu S (2020) Simula ion o he
s ong g ound mo ion o he 20 July 2017 (Mw. 6.6) Bod-
um-Kos ea hquake. Bull Ea hq Eng 18(13):5807–5825
Tsioulou A, Ta lanidis AA, Galasso C (2019) Valida ion o
s ochas ic g ound mo ion model modi ica ion by compa i-
son o seismic demand o eco ded g ound mo ions. Bull
Ea hq Eng 17(6):2871–2898 (Sp inge Ne he lands)
VahidiFa d H, Za a ani H, Sabbagh-Yazdi SR e al (2017) Seis-
mic haza d analysis using simula ed g ound-mo ion ime
his o ies: he case o he Se id ud dam, I an. Soil Dyn
Ea hq Eng 99:20–34
Wa son-Lamp ey J, Ab ahamson N (2006) Selec ion o g ound
mo ion ime se ies and limi s on scaling. Soil Dyn Ea hq
Eng 26(5):477–482
Wells DL, Coppe smi h KJ (1994) New empi ical ela ionships
among magni ude, up u e leng h, up u e wid h, up u e
a ea, and su ace displacemen . Bull Seismol Soc Am
84(4):974–1002 (GeoScienceWo ld)
Wi he s KB, Mosche i MP, Thompson EM (2020) A machine
lea ning app oach o de eloping g ound mo ion mod-
els om simula ed g ound mo ions. Geophys Res Le
47(6):e2019GL086690 (John Wiley & Sons, L d)
Xu J, Feng DC (2018) Seismic esponse analysis o nonlin-
ea s uc u es wi h unce ain pa ame e s unde s ochas-
ic g ound mo ions. Soil Dyn Ea hq Eng 111:149–159
(Else ie )
Yaghmaei-Sabegh S, Hassani B (2020) In es iga ion o he
ela ion be ween Vs30 and si e cha ac e is ics o I an
based on ho izon al- o- e ical spec al a ios. Soil Dyn
Ea hq Eng 128:105899
Yamamo o Y, Bake JW (2013) S ochas ic model o ea h-
quake g ound mo ion using wa ele packe s. Bull Seismol
Soc Am 103(6):3044–3056
Publishe ’s No e Sp inge Na u e emains neu al wi h ega d
o ju isdic ional claims in published maps and ins i u ional
a ilia ions.
