Landslides
Rosa M. Palau · Ma c Be engue · Ma cel Hü limann · Daniel Sempe e‑To es
Applica ion o a uzzy e i ica ion amewo k
o he e alua ion o a egional‑scale landslide
ea ly wa ning sys em du ing he Janua y 2020
Glo ia s o m in Ca alonia (NE Spain)
Abs ac The Glo ia s o m ain alls a ec ed Ca alonia om 20 o
23 Janua y 2020 and igge ed mul iple landslides, some o which
a ec ed buildings and in as uc u es (such as oads and ailways).
This pape p esen s he ain all and landslide da ase s collec ed
du ing he e en , and e alua es he pe o mance o a egional land-
slide ea ly wa ning sys em (LEWS) du ing he Glo ia s o m apply-
ing a uzzy e i ica ion me hod. The majo i y o he in en o ied
landslides can be classi ied as slides, in ol ing a limi ed olume o
sedimen (up o 10 m3), and we e igge ed in cu slopes along lin-
ea in as uc u es. Rain all accumula ions we e signi ican in he
whole egion, especially in he Mon seny a ea, whe e o e 450 mm
we e egis e ed in 96 h. Gene ally, he LEWS compu ed mode a e
and high wa nings in he a eas whe e la ge ain all amoun s we e
eco ded, and showed good co espondence wi h he loca ions
whe e landslides we e epo ed. The uzzy e i ica ion me hod
has been applied using neighbou ing windows o di e en sizes
o ob ain scale-dependan in o ma ion on he LEWS pe o mance.
The skill o he LEWS conside ably imp o es when enla ging he
neighbou ing window size om 500 m o 1 km.
Keywo ds Glo ia s o m · MORLE · Quan i a i e ain all
es ima ion · Landslide ea ly wa ning sys em · Fuzzy e i ica ion ·
Unce ain y
In oduc ion
Mul iple-occu ence egional landslide e en s (MORLEs) a e de ined
as hund eds o indi idual landslides occu ing almos simul aneously
o e la ge a eas (C ozie 2005). Usually, MORLEs a e cons i u ed by
shallow slides o lows ha a e igge ed in s eep slopes by in ense
ains o ms o ea hquakes. MORLEs ha e been desc ibed in di e -
en egions a ound he globe, such as New Zealand (C ozie 2005),
Taiwan (Yu e al. 2006), China (Yang e al. 2020), he USA (Campbell
1975; Whi ake and McShane 2012), Swi ze land (Nicole e al. 2013),
o I aly (C os a and F a ini 2003; Lomba do e al. 2018).
Se e al MORLEs also happened in he egion o Ca alonia (NE
Spain) in he pas : Oc obe 1940 (Po illa 2014), Augus 1963 (Po illa
2014), No embe 1982 (Galla and Clo e 1988; Co ominas and Alonso
1990), June 2008 (Po illa e al. 2010), o June 2013 (Shu e al. 2019). These
MORLEs mainly a ec ed he Py enees and P e-Py enees and we e
associa ed wi h se e e ain all e en s and looding. Mos ecen ly, om
20 o 23 Janua y 2020, an ex ao dina y E-NE cyclonic s o m (named
Glo ia) a ec ed he egion o Ca alonia. The signi ican and widesp ead
Glo ia s o m ain alls igge ed mul iple landslides, especially in he
Mon seny (Fig. 1).
The high numbe o landslides and he la ge a ea a ec ed by
MORLEs no mally suppose a challenge o he au ho i ies in cha ge
o managing he isk and he main enance o oads and ailways.
In his con ex , egional landslide ea ly wa ning sys ems (LEWS)
may help o iden i y he ime and loca ion whe e landslides a e
mos likely o occu and inc ease hei p epa edness (Al ie i e al.
2012; UNISDR 2015).
In he las 20 yea s, egional landslide ea ly wa ning sys ems
ha e been de eloped co e ing mul iple egions, e.g. Sou he n Cali-
o nia (Baum and God 2010), Rio de Janei o (Cal ello e al. 2015),
Indonesia (Hidaya e al. 2019), Hong Kong (Lloyd e al. 2001), Japan
(Osanai e al. 2010), he Zhenjiang p o ince in China (Yin e al.
2008), he ci y o Busan in Sou h Ko ea (Pa k e al. 2019), No way
(K øgli e al. 2018), he Emilia-Romagna and Campania egions in
I aly (Piciullo e al. 2017b; Segoni e al. 2018), and Ca alonia in Spain
(Be engue e al. 2015; Palau e al. 2020). Usually, LEWS de e mine
he a eas ha a e p one o landslides employing suscep ibili y maps
and assess whe he a ain all e en migh igge a landslide using
ain all h esholds (Aleo i 2004; Guzze i e al. 2007; Papa e al.
2013; Rossi e al. 2017; Pan e al. 2018). The majo i y o egional-scale
LEWS use ain gauge da a o assess he ain all haza d. Howe e ,
in many cases, he densi y o ain gauge ne wo ks is low, and land-
slide- igge ing ain all ends o be unde es ima ed (Nikolopoulos
e al. 2014). O he LEWS use emo e sensing da a such as sa elli e o
g ound-based ada ain all p oduc s (Be engue e al. 2015; Rossi
e al. 2017; Ki schbaum and S anley 2018).
LEWS need egula and sys ema ic pe o mance analysis o
assu e he eliabili y o he models. Up o da e, esea ch has mainly
ocused on he alida ion and imp o emen o ain all h esholds
(Ga iano e al. 2015; B une i e al. 2018) and suscep ibili y maps
(Ki schbaum e al. 2016). Only a ew s udies ha e pu hei a en-
ion in back-analysing he ou pu wa nings and i s co espondence
wi h epo ed landslides. Cal ello and Piciullo (2016) and Piciullo
e al. (2020) p oposed he EDUMAP me hod o he e alua ion o
egional-scale LEWS du ing long pe iods. This me hodology con-
side s he possible occu ence o mul iple landslides, and akes
in o accoun he ela ion be ween he du a ion o he wa nings
and he landslide epo ing ime. Ki schbaum e al. (2009) and
Pa k e al. (2020) p oposed a neighbou ing window o de e mine
he pe o mance o global and egional-scale LEWS. In his line,
uzzy e i ica ion me hods ha e long been employed o assess he
pe o mance o mesoscale high- esolu ion p ecipi a ion o ecas s
(B ooks e al. 1998; A ge 2001; Dam a h 2004; Robe s and Lean
2008; Ebe 2008; Ma sigli e al. 2008; Cla k e al. 2010) and could be
applied o he e alua ion o LEWS pe o mance. Fuzzy e i ica ion
Landslides
DOI 10.1007/s10346-022-01854-2
O iginal Pape
Recei ed: 24 Feb ua y 2021
Accep ed: 4 Feb ua y 2022
© The Au ho (s) 2022
Landslides
O iginal Pape
Landslides
me hods analyse how he e alua ion esul s change when elaxing
he condi ion o co-localiza ion be ween simula ions and obse a-
ions (i.e. wa nings and landslide in en o y poin s).
Ha ing landslide in en o ies ha a e comple e in space and
ime is c ucial o es ablish eliable LEWS and o e alua e hei
pe o mance. His o ically, landslide in en o ies we e collec ed
ocusing on small a eas om he in e p e a ion o ae ial pho o-
g aphs, emo e sensing da a, ield su eys, and local epo s (Galli
e al. 2008; Guzze i e al. 2012). Al e na i ely, in en o y da a can
be ob ained om da a sou ces such as newspape s epo s, and
c owdsou cing (Guzze i e al. 1994; Ki schbaum e al. 2010; Ekke
e al. 2013; Juang e al. 2019). Howe e , hese in en o ies a e o en
incomple e and usually biased o landslides ha a ec ed u ban
a eas o in as uc u es (A dizzone e al. 2002).
The la ge numbe o landslides ha we e epo ed du ing he
Glo ia s o m gi es us a unique oppo uni y o analyse he pe o -
mance o he exis ing landslide ea ly wa ning sys em o he egion
o Ca alonia. To do so, we p opose o apply a uzzy e i ica ion
me hod widely employed o he e i ica ion o p ecipi a ion o e-
cas s using se e al neighbou ing window sizes. The objec i es o
he s udy a e (i) o analyse he Glo ia s o m ain all e en and he
landslides ha we e igge ed, and (ii) o assess he pe o mance
o he LEWS du ing he Glo ia s o m and deduce he ex en o he
LEWS skill using an in en o y which has spa ial and empo al
unce ain ies.
Se ings
Geology and clima e o Ca alonia
Ca alonia is loca ed in he NE o he Ibe ian Peninsula and
co e s an a ea o a ound 32,000 km2. I s o og aphy (Fig. 1a) is
he esul o (i) he o ma ion o he Py enees wi h peaks o e
3000 m asl., he Ca alan Coas al Range, and he Ibe ian Range
du ing he Paleogene; (ii) he la e deposi ion o sedimen s in
he Eb o Basin; and (iii) he o ma ion o a se ies o ho s and
g abens mo e o less pa allel o he coas line du ing he Miocene
(Be as egui e al. 2010). The bed ock li hology is e y di e se
and includes igneous, sedimen a y, and me amo phic ma e ials.
In many a eas, he bed ock is co e ed by su icial o ma ions
o a ied hickness. While in some a eas hese deposi s me ely
consis o a ew cen ime es, in o he s, he su icial o ma ions
can be e y hick, o he o de s o me es. The li hology o hese
su icial deposi s is also e y a iable and can be e y di e en
om one a ea o ano he .
Ca alonia’s clima e a ies, bu can be classi ied as Medi e anean
(Embe ge 1952). Nea he coas , he wea he is mild and empe a e,
wi h a mean annual empe a u e o 17 °C. Inland, he clima e is
con inen al wi h cold win e s, ho summe s, and less abundan
p ecipi a ion. The Py enees p esen a high-al i ude clima e wi h
abundan snow and empe a u es below 0 °C du ing win e . The
ainies seasons a e gene ally sp ing and au umn, excep o he
Py enees, whe e he ainies season is summe . The mean annual
ain all anges om less han 400 mm in some pa s o he Eb o
Basin o o e 1200 mm in he Py enees. In Ca alonia, he 10-yea
e u n pe iod 24-h ain all accumula ion commonly exceeds
100 mm (Cla e o e al. 1996). Daily accumula ions o o e 200 mm
can be egula ly seen a leas once a yea in he coas al a ea (Ma ín
Vide and Olcina Can os 2001). The Glo ia s o m was a a he
unusual e en o hea y ains du ing he d ies mon hs o he yea .
Landslides a e gene ally igge ed by ei he (i) con ec i e ain-
all e en s wi h high in ensi ies which cause widesp ead shallow
slides and deb is lows, ypical om mid-summe o ea ly au umn,
o (ii) long-las ing ain alls wi h mode a e in ensi ies which usually
igge o eac i a e ea h lows and mid-size slides, common du -
ing sp ing and au umn (Co ominas e al. 2002, 2015; Abancó e al.
2016). The Glo ia s o m ain alls happened du ing win e , bu s ill
igge ed a signi ican numbe o landslides.
Fig. 1 a Gene al o e iew map o Ca alonia. The g een diamonds
show he loca ion o he wea he ada s and yellow ci cles he 183
ain gauges. The ou ed ci cles show he loca ion o he Vilad au
(WS), PN dels Po s (X5), To oella de Flu ià (XZ), and Els Hos ale s de
Pie ola (CE) ain gauges. The ed dashed polygon po ays he loca-
ion o he Mon seny a ea. b Densi y map o he landslides igge ed
by he Glo ia s o m and ga he ed in he in en o y. The black c osses
ep esen he landslide poin s o he ICGC and he #Eslla ica in en-
o ies. The main i e s a e ep esen ed as blue lines. The loca ion o
Ba celona is indica ed wi h a black ci cle
Landslides
LEWS desc ip ion
He ein, we b ie ly desc ibe he LEWS o he egion o Ca alo-
nia. Mo e de ails can be ound in Be engue e al. (2015) and
Palau e al. (2020). The LEWS has been designed wi h he aim o
issuing eal- ime wa nings o he au ho i ies in cha ge o man-
aging landslide isk in Ca alonia. I is unning p e-ope a ionally
a he uni e si y se e s o es ing pu poses. The LEWS com-
bines wo inpu pa ame e s: (i) a 30-m- esolu ion suscep ibili y
map (Fig. 2a) and (ii) high- esolu ion ain all obse a ions. The
ou pu o he LEWS is upda ed e e y ime new ain all obse a-
ions a e a ailable and consis s on a map showing a quali a i e
wa ning le el.
The suscep ibili y map (Fig. 2a) is used o depic he loca ions
whe e landslides may occu . I was de i ed by Palau e al. (2020)
applying a uzzy logic me hodology o combine slope angle and
land use and land co e in o ma ion.
To assess i a ain all e en has he po en ial o igge ing
a landslide, he in ensi y–du a ion– equency (IDF) cu es o
he Fab a me eo ological obse a o y in Ba celona (Casas e al.
2004) a e used o de ine ou ain all haza d le els (Fig. 2b). To
sepa a e di e en ain all e en s, an in e -e en pe iod o 6 h
wi hou ain all is used. Nei he he an eceden ain all no soil
mois u e condi ions a e no aken in o accoun in he cu en
e sion o he LEWS.
Finally, he ain all haza d and he suscep ibili y a e com-
bined h ough a wa ning ma ix. The esul is a 30-m g idded
wa ning le el map. Each wa ning le el (‘ e y low’, ‘low’, ‘mode -
a e’, and ‘high’) indica es he possibili y ha a landslide is ig-
ge ed a a speci ic loca ion. Addi ionally, a summa y showing
he maximum wa ning le el compu ed wi hin he i s second-
and hi d-o de hyd ological sub-basins as de ined by S ahle
(1957) is p o ided.
Addi ional analysis o ecen ain all e en s ha igge ed
landslides in Ca alonia showed ha he ain all in ensi y–du a ion
(I-D) h esholds ini ially applied o de e mine he ‘Mode a e’ and
‘High’ wa ning le els we e oo low. The e o e, he e we ha e adap ed
he I-D h esholds employed in Palau e al. (2020). The 5-yea and
20-yea e u n pe iod I-D cu es ha e been used o de ine he
‘Mode a e’ and ‘High’ ain all haza ds espec i ely.
Desc ip ion o he Glo ia s o m
F om 20 o 23 Janua y 2020, he Glo ia s o m a ec ed he egion
o Ca alonia, causing se e al di e en haza ds such as s o m
su ges, e osion o beaches in coas al a eas, loods, and landslides.
Acco ding o he Ca alan O ice o he Clima e Change (Canals
and Mi anda 2020; OCCC 2020), he economic losses due o hese
impac s exceeded 500 million eu os. The Glo ia s o m was excep-
ional, because i ook place du ing win e , an unusual season o
o en ial ain alls in his a ea, and also because o i s long du a ion.
This sec ion i s p esen s he me eo ological si ua ion and
analyses he ain all accumula ions. Then, he landslides igge ed
by he Glo ia s o m landslide in en o ies a e desc ibed. Finally, he
LEWS Glo ia s o m ou pu s a e s udied.
Me eo ological si ua ion
The Glo ia s o m was p eceded by an an icyclone ha las ed o e
a mon h, du ing which i did no ain in Ca alonia. On 18 Janua y
2020, a cold on coming om he No h A lan ic en e ed h ough
he no h wes o he Ibe ian Peninsula and mo ed sou h owa ds
he Medi e anean Sea. On he B i ish Isles, an unusual an icyclonic
si ua ion eco ded p essu es up o 1050 hPa, he highes p essu e
since 1957 (Se ei Me eo ològic de Ca alunya 2020a). This high
p essu e had an elonga ed shape om eas o wes and co e ed a
la ge pa o cen al Eu ope.
The Glo ia s o m was he esul o he combina ion o he
unusual high p essu es on he B i ish Isles and he low p essu es
loca ed on he sou h o he Ibe ian Peninsula. The g adien o p es-
su es be ween hese wo cen es caused s ong eas -no heas winds
and, p o ided a high humidi y and abundan and widesp ead p e-
cipi a ion (Se ei Me eo ològic de Ca alunya 2020b). The du a ion
o he Glo ia s o m was long because he No h A lan ic cold-ai
mass was s a iona y o e Ca alonia o se e al days.
P ecipi a ion analysis
The ain all da ase s used in his s udy consis o he measu emen s
o 187 ipping bucke ain gauges om he Me eo ological Se ice
Fig. 2 Suscep ibili y map (a)
and ain all in ensi y-du a ion
h esholds (b) employed by
he LEWS
Landslides
O iginal Pape
Landslides
o Ca alonia (SMC), and he quan i a i e p ecipi a ion es ima es
(QPEs) om he composi e o he obse a ions o he SMC ada
ne wo k (XRAD). The loca ion o he ain gauges and he ada s is
po ayed in Fig. 1a.
Rada QPEs ha e been p oduced om he olume scans o
C eu del Ven , La Panadella, and Puig d’A ques C-band single-
pola isa ion Dopple ada s o he SMC wi h he in eg a ed ool
o hyd ome eo ological o ecas ing (EHIMI, Co al e al. 2009).
The EHIMI ool includes a chain o quali y con ol, co ec ion,
mosaicking, and accumula ion algo i hms o gene a e QPE p oduc s
om aw ada obse a ions. The p oduc used he e is he 30-min
p ecipi a ion accumula ion ield wi h a spa ial esolu ion o 1 km.
Rain gauge measu emen s and ada obse a ions ha e been
combined o ob ain an imp o ed QPE applying he me hod p o-
posed by Velasco-Fo e o e al. (2009) and Cassi aga e al. (2020).
This me hod employs a geos a is ics echnique known as k iging
wi h an ex e nal d i (KED) o in e pola e he ain gauge obse a-
ions using ada ain all as a seconda y a iable ha p o ides he
d i o he ain all ield be ween ain gauges. As shown by Velasco-
Fo e o e al. (2009), his me hod bene i s om he di ec su ace
ain all obse a ions o he ain gauges loca ed wi hin he s udy
a ea, and he ada desc ip ion o he spa io empo al a iabili y
o he ain all ield.
Figu e 3 p esen s he daily p ecipi a ion accumula ions om
20 o 23 Janua y 2020. The e olu ion o he Glo ia s o m and he
spa io empo al a iabili y o he ain all ield can be obse ed in
hese plo s. I also shows he loca ions o landslide epo s in ela-
ion o he ain all.
The s o m began on 20 Janua y 2020 when snow and ain we e
obse ed in he no heas and he sou h (Fig. 3a). On 21 Janua y
2020, p ecipi a ions ell o e he en i e egion. S ill, hey we e mo e
abundan pa allel o he coas line, whe e he 24-h ain all accu-
mula ions exceeded 200 mm a he Mon seny a ea and he Ibe ian
ange (Fig. 3b). Du ing 22 Janua y 2020, ain all ell in e mi en ly
o e mos o Ca alonia. Mo e han 140 mm we e accumula ed in he
Mon seny a ea (Fig. 3c). Addi ionally, impo an ain all accumula-
ions we e eco ded a he sou hwes o Ca alonia, he Py enees, and
he P e-Py enees. The main p ecipi a ion sys em mo ed owa ds
he no h du ing he mo ning o 23 Janua y 2020. Rain all ell in e -
mi en ly wi h mode a e and low in ensi ies. Al hough ain all accu-
mula ions we e no as ele an as he p e ious days (Fig. 3d), hey
we e s ill signi ican in he Mon seny a ea, whe e o e 100 mm we e
eco ded in some a eas, and in he P e-Py enees.
The o al accumula ed ain all du ing he 4 days was signi ican
o e mos o Ca alonia (Fig. 4). The la ges ain all amoun s ell
o e he no h-eas , wi h a ound 480 mm in he Mon seny a ea.
Du ing he i s day o he Glo ia s o m, no landslides we e
epo ed. In he ollowing days, he a eas ha eco ded he highes
ain all accumula ions coincide a he well wi h he places whe e
landslides we e epo ed (black ci cles in Figs. 3 and 4).
Analysis o he quali y o he p ecipi a ion es ima es
This sec ion p esen s an analysis o he quali y o he p ecipi a ion
es ima es ob ained applying he KED me hod. The pe o mance
has been e alua ed by lea e-one-ou c oss- alida ion using he
obse a ions a he ain gauges as he e e ence. To do so, we ha e
applied he KED me hod emo ing one o he ain gauges om he
calcula ion o es ima e he ain all a he loca ion o he emo ed
ain gauge. Then, we ha e compa ed he es ima ed alue wi h he
obse ed ain all. This p ocess has been epea ed o e e y 30 min
and each o he 187 conside ed ain gauges.
Figu e 5 shows he compa ison be ween he o al p ecipi a ion
accumula ions du ing he e en ob ained om c oss- alida ion
and he o al accumula ions obse ed a each o he ain gauges.
Addi ionally, ou s a is ics ha e been added o he sca e plo : he
bias, he s anda d de ia ion o he e o (SD e o ), he oo mean
squa ed e o (RMSE), and he oo mean squa ed ela i e e o
(RMSR). The e en KED es ima es gene ally show a good ag ee-
men wi h he e en accumula ions eco ded a ain gauges. The SD
o he e o and he RMSE a e simila , a ound 31 mm; he e o e, he
bias is a he low. And he e en RMSR is 22%.
The esul s om he compa ison o he hye og aphs ob ained
by c oss- alida ion and he hye og aphs om ain gauge obse a-
ions o he ou selec ed ain gauges dis ibu ed o e he Ca a-
lan e i o y (see ed ci cles in Fig. 1) a e p esen ed in Fig. 6. The
e olu ion o 30-min accumula ions ep oduces he obse a ions
sa is ac o ily a he majo i y o he ain gauges. Howe e , in some
loca ions (e.g. Vilad au and PN dels Po s), he KED unde es ima es
he measu ed in ensi ies. In o he si es, such as To oella de Flu ià,
he KED sligh ly o e es ima es he obse ed ain all. The esul s o
he 30-min accumula ions ob ained a all he a ailable ain gauges
show ha he RMSE anges be ween 0.15 and 1.72 mm. In he calcu-
la ion o he RMSR, we ha e imposed a h eshold o 1 mm/30 min,
and he esul s o RMSR ange be ween 21.7 and 107.4%, wi h a
median alue o 42.8%. The e o s in small accumula ions ha e a
signi ican e ec in he calcula ion o RMSR, and he la ge alues
a e ob ained in a eas wi h e en accumula ions be ween 100 and
150 mm.
The esul s p esen ed in his sec ion quan i a i ely desc ibe he
unce ain y in he QPEs ob ained by KED du ing he Glo ia s o m.
These QPEs a e he p ecipi a ion inpu s o he Ca alonia egion
LEWS and, he e o e, hei unce ain y will a ec he pe o mance
o he LEWS and he quali y o he wa nings du ing his e en (see
he “Analysis o he LEWS ou pu s” sec ion).
Landslide in en o y and impac s
The signi ican ain all accumula ions and high in ensi ies egis-
e ed du ing he Glo ia s o m igge ed a la ge numbe o land-
slides o e di e en a eas in Ca alonia. One o he main challenges
o he e alua ion o he pe o mance o LEWS is he a ailabili y
o a comple e landslide in en o y. In Ca alonia, no sys ema ic and
o icial landslide in en o y exis s. The e o e, in his s udy we ha e
used in o ma ion con ained in wo di e en landslide in en o ies:
he in en o y o he Ca og aphic and Geological Ins i u e o Ca a-
lonia (ICGC in en o y; González e al. 2020), and he #Eslla ica
in en o y.
The ICGC in en o y ga he s landslide in o ma ion om se e al
sou ces such as epo s om di e en adminis a ions (municipali-
ies, coun y councils, ci il p o ec ion, moun ain ange s, and o he
ins i u ions), in e p e a ion o ae ial pho og aphs aken a e he
Glo ia s o m along some i e banks, and media epo s. I includes
a o al o 348 en ies. Howe e , in o ma ion o hese 348 landslides
is no comple e and many imes lacks de ails. Fo example, he
ICGC in en o y does no include olume in o ma ion. The majo i y
Landslides
o landslides a e classi ied acco ding o he Va nes (1978) classi ica-
ions. Ye , some epo s may be due o accumula ion o sedimen
on oads associa ed wi h o he p ocesses such as wa e e osion. All
he en ies o he ICGC in en o y include in o ma ion on he da e.
Howe e , some en ies ha e no clea da e and he day o occu ence
du ing he Glo ia s o m is he e o e unce ain. Addi ionally, he
loca ion o a ound 25% o he epo s is unce ain, and 23 landslides
a e loca ed in u ban a eas in la lands, whe e no slope o alus could
be obse ed in hei icini y. The e o e, hese poin s ha e no been
used o ou analysis since we conside ed hei spa ial unce ain y
was oo la ge.
The #Eslla ica in en o y collec s da a om social ne wo k
pos s o local obse e s. A o al o 108 geoloca ed landslides we e
epo ed h ough social ne wo ks and ha e been included in he
#Eslla ica in en o y. The majo i y o #Eslla ica landslide epo s
included a pho og aph o ideo o he ini ia ion o deposi a ea
(see examples in Fig. 7). Mos o he landslide loca ions ha e been
checked by p e-s o m Google S ee View. Using his in o ma ion,
Fig. 3 Daily ain all accumula ions du ing he Glo ia s o m: a 20
Janua y 2020, b 21 Janua y 2020, c 22 Janua y 2020, d 23 Janua y
2020. Black ci cles ep esen he landslides included in he in en o y
each day. In he ollowing sec ions, mo e de ails abou he in en o y
and he landslides a e gi en
Landslides
O iginal Pape
Landslides
oge he wi h he desc ip ions p o ided in some pos s, he land-
slides ha e been classi ied in o di e en ypes acco ding o he
classi ica ions p oposed by Va nes (1978) and Hung e al. (2014).
Addi ionally, a measu e o he e en size has been assigned o each
in en o y en y o di e en ia e be ween h ee olume anges: less
han 1 m3, be ween 1 and 10 m3, and mo e han 10 m3. Since #Eslla -
ica epo s we e made by popula ion, mos landslides happening
o e nigh we e in o med du ing he mo ning. Some o he epo s
we e made once he s o m had ceased, and he e o e, he p ecise
igge ing da e is unce ain.
Fo his s udy, he ICGC and he #Eslla ica in en o ies ha e
been me ged, and duplica ed poin s ha e been emo ed. The inal
landslide da a se con ains 108 poin s om he #Eslla ica in en-
o y and 275 om he ICGC in en o y, esul ing in a o al o 383
landslide poin s.
The Mon seny is he a ea whe e he la ges densi y o landslides
was obse ed, 0.28 landslide/km2 (Fig. 1b). This densi y is a he
low, compa ed wi h he densi y o landslides obse ed o his o ical
MORLEs in Ca alonia (e.g. 1.5 landslides/km2 in he Py enees 1982
(Co ominas and Alonso 1990), and 1.16 landslides/km2 in Val d’A an
2008 (Shu e al. 2019). The di e ences may be pa ly due o he
comple eness o his o ical in en o ies, which ully co e ed smalle
egions inside Ca alonia wi h ield su eys and he in e p e a ion
o ae ial pho og aphs. This was no possible o he Glo ia s o m
in en o y due o he much la ge ex ension and because no pos -
e en ligh su eillances we e made o e he mos a ec ed a eas.
The cha ac e is ics o he landslides igge ed by he Glo ia
s o m and con ained in he inal in en o y a e desc ibed he eunde .
The accumula ed ain all a he loca ion o he epo ed landslides
has been checked (Fig. 4). F om he 383 landslides used o his
s udy, mo e han 100 we e epo ed in places ha egis e ed e en
ain all accumula ions o e 300 mm in 96 h (Fig. 8a).
Acco ding o i s ype, 69% o he in en o ied landslides we e
slides, 20% alls, and 3% lows (Fig. 8b). The ype o he emaining
8% o e en s igge ed by he Glo ia s o m is unclea . The in en-
o ies do no ha e enough in o ma ion o de e mine wha ype o
ma e ial o sliding mechanism was mo e p edominan du ing he
Glo ia s o m. Rega ding he landslide olume, only in o ma ion
om he #Eslla ica epo s is a ailable. Fou een pe cen o he
landslides con ained in he #Eslla ica in en o y we e small, wi h
a olume o less han 1 m3. The olume o 56% o he #Eslla ica
landslide epo s anged be ween 1 and 10 m3. Only 12% o he
#Eslla ica epo s co espond o landslides wi h olumes la ge
han 10 m3. The olume o 18% o he en ies could no be es ab-
lished because no enough in o ma ion was gi en in he epo .
Fig. 4 Accumula ed ain all
om 20 Janua y 2020 00:00
o 23 Janua y 2020 24:00. The
black ci cles ep esen he
landslides included in he ICGC
and he #Eslla ica in en o ies
Landslides
The 5-m- esolu ion DEM has been used o es ima e he slope
angles, and he 30-m- esolu ion DEM has been employed o
ob ain he o ien a ion (ICGC 2013). Simila ly, land use and land
co e wi h a esolu ion o 30 m (MCSC-4, (CREAF 2009)) and
he g aph o he Ca alonia in as uc u es ne wo k (DGMT 2019)
ha e been applied o analyse he mos common land use and land
co e classes a he landslide loca ions and he p oximi y o oads
and ailway lines.
The majo i y o landslides we e loca ed a s eep slopes o o e
20° (Fig. 8c). A ound 27% o he e en s we e epo ed in slopes
wi h angles be ween 10 and 20°, and abou 16% in gen le slopes
wi h slope angles less han 10°. Such low slope angles a e a he
di icul o jus i y om a geo echnical poin o iew and may be
ela ed o spa ial unce ain y. No clea end can be obse ed
in he o ien a ion o he slopes whe e landslides we e epo ed.
Howe e , he o al numbe o e en s igge ed in eas , no h-eas ,
and sou h-eas acing slopes is sligh ly la ge han he sum o
he e en s a sou h, sou h-wes , and wes acing slopes (Fig. 8d).
The main wind di ec ion o he Glo ia s o m was owa ds wes -
no h wes ; hus, eas and sou h-eas acing slopes would be he
mos exposed.
Landslides mos equen ly occu ed in o es a eas (Fig. 8e) and
55 e en s we e epo ed in a eas wi h in as uc u es o buildings.
Two o he landslides con ained in he in en o y we e loca ed in
wa e bodies, which migh be ela ed o he scou ing in i e banks.
Mos o he epo ed landslides we e spo ed close o linea in a-
s uc u es (Fig. 8 ). A ound 64% we e igge ed be ween 0 and 10 m
away om he oad o ailway axis. The numbe o landslide epo s
diminishes wi h he dis ance om linea in as uc u es. Only 38%
o he epo ed landslides we e loca ed u he han 200 m. These
esul s p o ide wo conclusions: (i) mo e han hal o he epo s
we e ela ed o slope ailu es o oad cu s and embankmen s in
linea in as uc u es, and (ii) landslides happening in emo e
inhabi ed a eas may gene ally be un epo ed.
Analysis o he LEWS ou pu s
The LEWS has been un om 20 o 23 Janua y 2020 using he
KED 30-min ain all accumula ion es ima es as inpu s. Since he
landslide in en o y only has in o ma ion on he epo ing da e,
he co espondence be ween he 30-min wa ning ou pu s and he
landslide epo s could no be s udied. He e he daily maximum
wa ning le el has been used o analyse he quali y o he wa n-
ings compu ed each day o he Glo ia s o m.
Figu e 9 shows he sub-basin maximum wa ning le el sum-
ma y o each o he days o he Glo ia s o m and he posi ions
o in en o y epo s. F om he compa ison o he wa ning maps
o Fig. 9 and he 24-h ain all accumula ions o Fig. 3, i can be
obse ed ha gene ally, ‘Mode a e’ and ‘High’ wa nings we e
ob ained in he a eas ha eco ded he mos signi ican ain all
accumula ions du ing he co esponding day.
Gene ally, landslides (displayed as black ci cles in Fig. 9) we e
epo ed in places whe e he sub-basin daily wa ning summa y
was ‘Mode a e’ o ‘High’. A he eas e n hal o Ca alonia, ‘High’
wa nings we e displayed o e he a ea whe e he in en o y has
he highes densi y o landslides (Fig. 1b). ‘Mode a e’ and ‘High’
wa nings we e ob ained o e he sou h-wes o Ca alonia on 21
Janua y and o e he no h-wes o Ca alonia on 22 Janua y 2020,
bu ew landslides we e epo ed in hese a eas (Fig. 9b, c). The
Py enees, P e-Py enees, Ibe ian Range, and he wes e n Ca alan
Coas al Ranges ha e a low popula ion densi y. The e o e, i may
be he case ha some landslides migh ha e been un epo ed.
E alua ion o he pe o mance o he LEWS du ing he Glo ia
s o m
E alua ing he pe o mance o a high- esolu ion LEWS o e he
en i e Ca alonia is challenging because o he spa ial and empo-
al unce ain ies o he landslide in en o y as well as i s incom-
ple eness (see he “Landslide in en o y and impac s” sec ion).
T adi ional e i ica ion me hods ma ch he loca ion and ime
o he wa nings o he p ecise loca ion and ime o he epo ed
landslides o analyse he pe o mance o a LEWS. Consequen ly,
he unce ain ies and incomple eness o landslide in en o ies
ha e an e ec on he esul s o adi ional e i ica ion me hods.
Fuzzy e i ica ion me hods a e an al e na i e ha does no
equi e he exac coincidence be ween wa nings and obse a-
ions. Ins ead, such me hods assume ha he loca ion and ime
o wa nings can be close o he loca ion and ime o landslide
obse a ions bu s ill be use ul (Ebe 2008, 2009). To do so, uzzy
e i ica ion me hods look in space– ime neighbou ing windows
a ound each obse ed e en o e alua e he pe o mance o he
model, g an ing some lexibili y in he ma ching be ween he
p edic ion and he obse a ion.
Fuzzy e i ica ion echniques ha e been widely applied o
measu e he pe o mance o high- esolu ion mesoscale p ecipi a-
ion o ecas s (Dam a h 2004; Theis e al. 2005; Ebe 2008; Cla k
e al. 2010). In many cases, uzzy e i ica ion me hods analyse
Fig. 5 C oss- alida ion sca e plo compa ing he obse ed o al
accumula ed ain all a each o he 187 ain gauges (R) and he KED
es ima ed alue om he ada obse a ions (G)
Landslides
O iginal Pape
Landslides
he e ec o a ying he size o he neighbou ing windows. The
esul s can be exploi ed o de e mine a which scales he o ecas s
should be used o mee ce ain accu acy equi emen s (e.g. Ebe
2008, 2009). In his sec ion, we ha e applied a uzzy e i ica ion
me hod ha is used in me eo ology o e alua e he pe o mance
o he Ca alonia egion LEWS o he Glo ia s o m pe iod, and
deduce he scales o which he wa nings a e eliable.
Desc ip ion o he e i ica ion me hod
The uzzy e i ica ion me hod ha has been applied o he e alu-
a ion o he 30-m- esolu ion wa nings du ing he Glo ia s o m is
known as he ‘minimum co e age c i e ion’ (Dam a h 2004; Ebe
2008). This me hod supposes ha he loca ion and ime o he
obse a ions a e co ec , and conside s a neighbou ing window
Fig. 6 Obse ed (black line)
and es ima ed (blue line)
hye og aphs om c oss-
alida ion o ou ain gauges
om 20 Janua y 2020 00:00
o 23 Janua y 2020 24:00. The
loca ion o he ain gauges
can be obse ed in Fig. 1.
The ime s ep is 30 min. We
ha e imposed a 1 mm/30 min
h eshold o he compu a ion
o he mean ela i e e o s.
The e o e, he ela i e e o has
been compu ed o ime-s eps
when he obse a ions meas-
u ed a leas 1 mm/30 min.
G s a es o he o al e en
ain all accumula ion (Accum)
eco ded a each ain gauge.
R e e s o he es ima ed o al
e en ain all accumula ion a
each ain gauge loca ion
Landslides
a ound each obse a ion o sea ch o wa nings. The minimum
co e age c i e ion me hod assumes e en s a e equally likely o
occu anywhe e wi hin he neighbou ing window. Then, ca ego ical
sco es based on he con ingency able a e employed o he e i ica-
ion o he wa nings (Fawce 2006).
Fo he e i ica ion pu poses, we ha e conside ed ha a wa n-
ing has been gi en when he wa ning le el was ei he ‘Mode a e’ o
‘High’, and no wa ning has been gi en when he wa ning le el was
ei he ‘Low’ o ‘Ve y Low’. The landslides con ained in he in en-
o y ha e been used as e e ence. Following he minimum co e age
c i e ion, he pixels ha all inside he neighbou ing windows a e
used o assess he ue posi i es and he alse nega i es (Fig. 10): A
ue posi i e is an ou come whe e he LEWS co ec ly displays a
leas one wa ning wi hin he neighbou ing window o a landslide
obse a ion. In con as , a alse nega i e is an ou come whe e he
LEWS inco ec ly displays no wa ning wi hin he neighbou ing
Fig. 7 Examples o landslides
igge ed by he Glo ia s o m
in he Mon seny a ea. a Ro a-
ional slide in a collu ium slope
(pho o cou esy o Clàudia
Abancó). b Ro a ional slide ha
a ec ed a oad embankmen
and pa s o na u al slopes
(pho o cou esy o Roge Ruiz)
Fig. 8 His og ams showing he dis ibu ion o Glo ia s o m land-
slide epo s con ained in he ICGC and he #Eslla ica in en o ies
acco ding o a ain all accumula ion, b landslide ype, c slope angle,
d o ien a ion, e land use and land co e , and dis ance o he closes
oad o ailway line axis
Landslides
O iginal Pape
Landslides
(DRS-01–2015–700099) and he Spanish na ional p ojec s SMuCPhy
and EROSLOP (BIA 2015–67500-R and PID2019–104266RB–I00/
AEI/).
Decla a ions
Con lic o in e es 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 Commons
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 Com-
mons 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 Com-
mons 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 eco mmons. o g/ licen ses/ by/4. 0/.
Re e ences
Abancó C, Hü limann M, Moya J, Be engue M (2016) C i ical ain all con-
di ions o he ini ia ion o o en ial lows. Resul s om he Rebaix-
ade ca chmen (Cen al Py enees). J Hyd ol 541:218–229. h ps:// doi.
o g/ 10. 1016/j. jhyd ol. 2016. 01. 019
Aleo i P (2004) A wa ning sys em o ain all-induced shallow ailu es.
Eng Geol 73:247–265
Al ie i L, Salamon P, Pappenbe ge F e al (2012) Ope a ional ea ly wa n-
ing sys ems o wa e - ela ed haza ds in Eu ope. En i on Sci Policy
21:35–49. h ps:// doi. o g/ 10. 1016/j. en sci. 2012. 01. 008
A dizzone F, Ca dinali M, Ca a a A e al (2002) Impac o mapping e o s
on he eliabili y o landslide haza d maps. Na Haza d 2:3–14. h ps://
doi. o g/ 10. 5194/ nhess-2- 3- 2002
A ge F (2001) Ve i ica ion o in ense p ecipi a ion o ecas s om single
models and ensemble p edic ion sys ems. Nonlinea P ocess Geophys
8:401–417. h ps:// doi. o g/ 10. 5194/ npg-8- 401- 2001
Baum RL, God JW (2010) Ea ly wa ning o ain all-induced shallow land-
slides and deb is lows in he USA. Landslides 7:259–272. h ps:// doi.
o g/ 10. 1007/ s10346- 009- 0177-0
Be as egui X, Casas JM, Liesa M e al (2010) His ò ia geològica de Ca a-
lunya. A les geològic de Ca alunya 68–77
Be engue M, Sempe e-To es D, Hü limann M (2015) Deb is- low o e-
cas ing a egional scale by combining suscep ibili y mapping and
ada ain all. Na Haza ds Ea h Sys Sci 15:587–602. h ps:// doi. o g/
10. 5194/ nhess- 15- 587- 2015
B ooks HE, Kay M, Ha JA (1998) Objec i e limi s on o ecas ing skill o
a e e en s. 19 h Con . Se e e Local S o ms. Ame ican Me eo ological
Socie y, Minneapolis, MN, USA, pp 552–555
B une i M, Massimo M, Pe uccacci S e al (2018) How a a e we om he
use o sa elli e ain all p oduc s in landslide o ecas ing? Remo e Sens
En i on 210:65–75. h ps:// doi. o g/ 10. 1016/j. se. 2018. 03. 016
Cal ello M, D’O si RN, Piciullo L e al (2015) The Rio de Janei o ea ly wa n-
ing sys em o ain all-induced landslides: analysis o pe o mance o
he yea s 2010–2013. In J Disas e Risk Reduc 12:3–15. h ps:// doi.
o g/ 10. 1016/j. ijd . 2014. 10. 005
Cal ello M, Piciullo L (2016) Assessing he pe o mance o egional
landslide ea ly wa ning models: he EDuMaP me hod. Na Haza d
16:103–122. h ps:// doi. o g/ 10. 5194/ nhess- 16- 103- 2016
Campbell RH (1975) Soil slips, deb is lows, and ains o ms in he San a
Monica Moun ains and icini y, Sou he n Cali o nia. US Geological
Su ey P o essional Pape 851, p 51
Canals M, Mi anda J (2020) Sob e el empo al Glo ia (19–23.01.20), els
seus e ec es sob e el país i el que se’n de i a: Repo de Respos a
Rapida (R3), 1a edició. ins i u d’Es udis Ca alans. Secció de Ciències
i ecnologia, Ba celona
Casas MC, Codina B, Redano A, Lo en e J (2004) A me hodology o
classi y ex eme ain all e en s in he wes e n Medi e anean
a ea. Theo e Appl Clima ol 77:139–150. h ps:// doi. o g/ 10. 1007/
s00704- 003- 0003-x
Cassi aga E, Gómez-He nández JJ, Be engue M e al (2020) Spa io empo al
p ecipi a ion es ima ion om ain gauges and me eo ological ada using
geos a is ics. Ma h Geosci. h ps:// doi. o g/ 10. 1007/ s11004- 020- 09882-1
Cla k AJ, Gallus WA, Weisman ML (2010) Neighbo hood-based e i i-
ca ion o p ecipi a ion o ecas s om con ec ion-allowing NCAR
WRF model simula ions and he ope a ional NAM. Wea he Fo ecas
25:1495–1509. h ps:// doi. o g/ 10. 1175/ 2010W AF222 2404.1
Cla e o P, Ma ín-Vide J, Raso-Nadal J (1996) A les Climà ic de Ca alunya.
Te moplu iome ia i RadiacióSola ; Ba celona, Spain, Spain
Co ominas J, Alonso E (1990) Geomo phological e ec s o ex eme
loods (No embe 1982) in he sou he n Py enees. In: Hyd ology in
moun ainous egions. II - A i icial Rese oi s; Wa e and Slopes. IAHS
no. 194, Lausanne, Swi ze land, pp 295–302
Co ominas J, Moya J, Hü limann M (2002) Landslide ain all igge s in
he Spanish Eas e n Py enees. In: 4 h EGS Plinius Con e ence “Medi e -
anean S o ms.” Edi ice, Mallo ca, Spain, pp 1–4
Co ominas J, Ma eos RM, Remondo J (2015) Re iew o landslide occu ence
in Spain and i s ela ion o clima e. In: Ken H, Lacasse S, Pica elli L (eds)
Slope sa e y p epa edness o impac o clima e change. Wo kshop on
Slope Sa e y P epa edness o E ec s o Clima e Change. Taylo & F an-
cis, Napoli, I aly, pp 1–28378
Co al C, Velasco D, Fo cadell D e al (2009) Ad ances in ada -based lood
wa ning sys ems. The EHIMI sys em and he expe ience in he Besòs lash-
lood pilo basin. Flood Risk Managemen Resea ch and P ac ice Ex ended
Abs ac s Volume 332 Pages Full Pape CDROM 1772, pp 1295–1303
CREAF (2009) Mapa de Cobe es del Sòl de Ca alunya (MSC-4), V4 edn
C os a GB, F a ini P (2003) Dis ibu ed modelling o shallow landslides
igge ed by in ense ain all. Na Haza ds Ea h Sys Sci 3:81–93.
h ps:// doi. o g/ 10. 5194/ nhess-3- 81- 2003
C ozie MJ (2005) Mul iple-occu ence egional landslide e en s in New
Zealand: haza d managemen issues. Landslides 2:247–256. h ps://
doi. o g/ 10. 1007/ s10346- 005- 0019-7
Dam a h U (2004) Ve i ica ion agains p ecipi a ion obse a ions o a
high densi y ne wo k – wha did we lea n? In: In e na ional Ve i ica-
ion Me hods Wo kshop. Mon eal, Canada
DGMT DG d’In aes uc u es de MT (2019) G a d’In aes uc u es Vec o-
ial de la DGIMT. Ba celona
Ebe EE (2008) Fuzzy e i ica ion o high- esolu ion g idded o ecas s:
a e iew and p oposed amewo k. Me eo ol Appl 15:51–64. h ps://
doi. o g/ 10. 1002/ me . 25
Ebe EE (2009) Neighbo hood e i ica ion: a s a egy o ewa ding close
o ecas s. Wea he Fo ecas 24:1498–1510. h ps:// doi. o g/ 10. 1175/
2009W AF222 2251.1
Ekke R, K æ ne K, Os A, Hums ad T (2013) egObs - public da abase o
submi ing and sha ing obse a ions. In: In e na ional Snow Science
Wo kshop. G enoble, Chamonix Mon -Blanc, p 5pp
Embe ge L (1952) Su le quo ien plu io he mique. C.R. Acade´mie Sci-
ence: Pa is; 2508–2510. Comp es Rendus De L’academie Des Sciences
234:2508–2510
Fawce T (2006) An in oduc ion o ROC analysis. Pa e n Recogn Le
27:861–874
Galla F, Clo e N (1988) Some aspec s o he geomo phic p ocesses
igge ed by an ex eme ain all e en : he No embe 1982 lood in
Eas e n Py enees. Ca ena Supplemen 79–95
Galli M, A dizzone F, Ca dinali M e al (2008) Compa ing landslide in en-
o y maps. Geomo phology 94:268–289. h ps:// doi. o g/ 10. 1016/j.
geomo ph. 2006. 09. 023
Landslides
Ga iano SL, B une i M, Io ine G e al (2015) Calib a ion and alida-
ion o ain all h esholds o shallow landslide o ecas ing in Sicily,
sou he n I aly. Geomo phology 228:653–665. h ps:// doi. o g/ 10.
1016/j. geomo ph. 2014. 10. 019
González M, Pinyol J, Micheo MJ e al (2020) El empo al Glo ia (19–
23/01/2020): Els e ec es dels p ocesos geológics sob e el e i o i.
Ins i u Ca og à ic i Geològic de Ca alunya, Ba celona
Guzze i F, Ca dinali M, Reichenbach P (1994) The AVI p ojec : a bib-
liog aphical and a chi e in en o y o landslides and loods in I aly.
En i on Manage 18:623–633. h ps:// doi. o g/ 10. 1007/ BF024 00865
Guzze i F, Pe uccacci S, Rossi M, S a k CP (2007) Rain all h esholds o
he ini ia ion o landslides in cen al and sou he n Eu ope. Me eo ol
A mos Phys 98:239–267. h ps:// doi. o g/ 10. 1007/ s00703- 007- 0262-7
Guzze i F, Mondini AC, Ca dinali M e al (2012) Landslide in en o y
maps: new ools o an old p oblem. Ea h Sci Re 112:42–66. h ps://
doi. o g/ 10. 1016/j. ea sc i e . 2012. 02. 001
Hidaya R, Su an o SJ, Hidayah A e al (2019) De elopmen o a landslide
ea ly wa ning sys em in Indonesia. Geosciences 9:451. h ps:// doi. o g/
10. 3390/ geosc ience s9100 451
Hung O, Le oueil S, Pica elli L (2014) The Va nes classi ica ion o land-
slide ypes, an upda e. Landslides 11:167–194. h ps:// doi. o g/ 10.
1007/ s10346- 013- 0436-y
ICGC (2013) Model d’Ele acions del Te eny de Ca alunya 5x5me es 1.0
(MET-5 1.0). ICGC
Juang CS, S anley TA, Ki schbaum DB (2019) Using ci izen science o
expand he global map o landslides: in oducing he Coope a i e
Open Online Landslide Reposi o y (COOLR). PLoS One 14:e0218657.
h ps:// doi. o g/ 10. 1371/ jou n al. pone. 02186 57
Ki schbaum DB, Adle R, Hong Y, Le ne -Lam A (2009) E alua ion o a p elimi-
na y sa elli e-based landslide haza d algo i hm using global landslide in en-
o ies. Na Haza d 9:673–686. h ps:// doi. o g/ 10. 5194/ nhess-9- 673- 2009
Ki schbaum DB, Adle R, Hong Y e al (2010) A global landslide ca alog
o haza d applica ions: me hod, esul s, and limi a ions. Na Haza ds
52:561–575. h ps:// doi. o g/ 10. 1007/ s11069- 009- 9401-4
Ki schbaum DB, S anley T, Ya heend adas S (2016) Modeling land-
slide suscep ibili y o e la ge egions wi h uzzy o e lay. Landslides
13:485–496. h ps:// doi. o g/ 10. 1007/ s10346- 015- 0577-2
Ki schbaum DB, S anley T (2018) Sa elli e-based assessmen o ain all-
igge ed landslide haza d o si ua ional awa eness. Ea h’s Fu u e
6:505–523. h ps:// doi. o g/ 10. 1002/ 2017E F0007 15
K øgli IK, De oli G, Colleuille H e al (2018) The No wegian o ecas ing
and wa ning se ice o ain all- and snowmel -induced landslides. Na
Haza d 18:1427–1450. h ps:// doi. o g/ 10. 5194/ nhess- 18- 1427- 2018
Lloyd DM, Wilkinson PL, O hmann MA, Ande son MG (2001) P edic ing
landslides: assessmen o an au oma ed ain all based landslide wa n-
ing sys em. In: KKS H, KS L (eds) 14 h Sou h Eas Asia Geo hecnical
Con e ence. Balkema, Hong Kong, pp 135–139
Lomba do L, Opi z T, Huse R (2018) Poin p ocess-based modeling o
mul iple deb is low landslides using INLA: an applica ion o he 2009
Messina disas e . S och En Res Risk Assess 32:2179–2198. h ps:// doi.
o g/ 10. 1007/ s00477- 018- 1518-0
Ma sigli C, Mon ani A, Paccangnella T (2008) A spa ial e i ica ion
me hod applied o he e alua ion o high- esolu ion ensemble o e-
cas s. Me eo ol Appl 15:125–143. h ps:// doi. o g/ 10. 1002/ me . 65
Ma ín Vide F, Olcina Can os J (2001) Climas y iempos de España. Alianza
Edi o ial
Nicole P, Fo es i L, Caspa O, Jaboyedo M (2013) Shallow landslide’s
s ochas ic isk modelling based on he p ecipi a ion e en o Augus
2005 in Swi ze land: esul s and implica ions. Na Haza d 13:3169–
3184. h ps:// doi. o g/ 10. 5194/ nhess- 13- 3169- 2013
Nikolopoulos EI, C ema S, Ma chi L e al (2014) Impac o unce ain y
in ain all es ima ion on he iden i ica ion o ain all h esholds o
deb is low occu ence. Geomo phology 221:286–297. h ps:// doi.
o g/ 10. 1016/j. geomo ph. 2014. 06. 015
OCCC (2020) L’impac e de la empes a Glo ia. In o me in e n, Ba celona
Osanai N, Shimizu T, Ku amo o K e al (2010) Japanese ea ly-wa ning o deb is
lows and slope ailu es using ain all indices wi h Radial Basis Func ion Ne -
wo k. Landslides 7:325–338. h ps:// doi. o g/ 10. 1007/ s10346- 010- 0229-5
Palau RM, Hü limann M, Be engue M, Sempe e-To es D (2020) In luence o
he mapping uni o egional landslide ea ly wa ning sys ems: compa ison
be ween pixels and polygons in Ca alonia (NE Spain). Landslides 17:2067–
2083. h ps:// doi. o g/ 10. 1007/ s10346- 020- 01425-3
Pan HL, Jiang YJ, Wang J, Ou GQ (2018) Rain all h eshold calcula ion o
deb is low ea ly wa ning in a eas wi h sca ci y o da a. Na Haza d
18:1395–1409. h ps:// doi. o g/ 10. 5194/ nhess- 18- 1395- 2018
Papa MN, Medina V, Cie o F, Ba eman A (2013) De i a ion o c i ical
ain all h esholds o shallow landslides as a ool o deb is low ea ly
wa ning sys ems. Hyd ol Ea h Sys Sci 17:4095–4107. h ps:// doi. o g/
10. 5194/ hess- 17- 4095- 2013
Pa k J-Y, Lee S-R, Lee D-H e al (2019) A egional-scale landslide ea ly
wa ning me hodology applying s a is ical and physically based
app oaches in sequence. Eng Geol 260:105193. h ps:// doi. o g/ 10.
1016/j. enggeo. 2019. 105193
Pa k J-Y, Lee S-R, Kim Y-T e al (2020) A egional-scale landslide ea ly
wa ning sys em based on he sequen ial e alua ion me hod: de el-
opmen and pe o mance analysis. Appl Sci 10:5788. h ps:// doi. o g/
10. 3390/ app10 175788
Piciullo L, Dahl M-P, De oli G e al (2017a) Adap ing he EDuMaP me hod
o es he pe o mance o he No wegian ea ly wa ning sys em o
wea he -induced landslides. Na Haza d 17:817–831. h ps:// doi. o g/
10. 5194/ nhess- 17- 817- 2017
Piciullo L, Ga iano SL, Melillo M e al (2017b) De ini ion and pe o mance o
a h eshold-based egional ea ly wa ning model o ain all-induced land-
slides. Landslides 14:995–1008. h ps:// doi. o g/ 10. 1007/ s10346- 016- 0750-2
Piciullo L, Ti an i D, Peco a o G e al (2020) S anda ds o he pe o -
mance assessmen o e i o ial landslide ea ly wa ning sys ems. Land-
slides 17:2533–2546. h ps:// doi. o g/ 10. 1007/ s10346- 020- 01486-4
Po illa M, Che alie G, Hü limann M (2010) Desc ip ion and analysis
o majo mass mo emen s occu ed du ing 2008 in he Eas e n Py -
enees. Na Haza ds Ea h Sys Sci 10:1635–1645. h ps:// doi. o g/ 10.
5194/ nhess- 10- 1635- 2010
Po illa M (2014) Recons ucción y análisis de ocu encias egionales de
múl iples e en os de mo imien os en masa gene ados po llu ias
his ó icas en los Pi ineos. In: TDX (Tesis Doc o als en Xa xa). Uni e si-
a Poli ècnica de Ca alunya
Robe s NM, Lean HW (2008) Scale-selec i e e i ica ion o ain all accu-
mula ions om high- esolu ion o ecas s o con ec i e e en s. Mon
Wea he Re 136:78–97. h ps:// doi. o g/ 10. 1175/ 2007M WR2123.1
Rossi M, Luciani S, Valigi D e al (2017) S a is ical app oaches o he
de ini ion o landslide ain all h esholds and hei unce ain y using
ain gauge and sa elli e da a. Geomo phology 285:16–27. h ps:// doi.
o g/ 10. 1016/j. geomo ph. 2017. 02. 001
Segoni S, Rosi A, Fan i R e al (2018) A egional-scale landslide wa ning
sys em based on 20 yea s o ope a ional expe ience. Wa e 10:1297.
h ps:// doi. o g/ 10. 3390/ w1010 1297
Se ei Me eo ològic de Ca alunya S (2020a) No a de P emsa. Balanç
d’una lle an ada his ò ica a Ca alunya. Ba celona, Spain
Se ei Me eo ològic de Ca alunya S (2020b) Bu lle í climà ic mensual.
gene del 2020. Ba celona, Spain
Shu H, Hü limann M, Molowny-Ho as R e al (2019) Rela ion be ween
land co e and landslide suscep ibili y in Val d’A an, Py enees (Spain):
his o ical aspec s, p esen si ua ion and o wa d p edic ion. Sci To al
En i on 693:133557. h ps:// doi. o g/ 10. 1016/j. sci o en . 2019. 07. 363
S ahle AN (1957) Quan i a i e analysis o wa e shed geomo phology.
EOS, T ansac ions - Ame ican Geophysical Union 38:913–920. h ps://
doi. o g/ 10. 1029/ 038 i006p 00913
Theis SE, Hense A, Dam a h U (2005) P obabilis ic p ecipi a ion o ecas s
om a de e minis ic model: a p agma ic app oach. Me eo ol Appl
12:257. h ps:// doi. o g/ 10. 1017/ S1350 48270 50017 63
UNISDR (2015) Sendai amewo k o disas e isk educ ion. In: Thi d
Uni ed Na ions Wo ld Con e ence on Disas e Risk Reduc ion, pp 1–25
Va nes DJ (1978) Slope mo emen s ypes and p ocesses. Landslides
analysis and con ol anspo a ion esea ch boa d. Na ional Acad-
emy o Sciences, Special Repo 176:11–33
Velasco-Fo e o CA, Sempe e-To es D, Cassi aga EF, Jaime Gómez-He nández
J (2009) A non-pa ame ic au oma ic blending me hodology o es ima e
Landslides
O iginal Pape
Landslides
ain all ields om ain gauge and ada da a. Ad Wa e Resou 32:986–
1002. h ps:// doi. o g/ 10. 1016/j. ad wa es. 2008. 10. 004
Whi ake KA, McShane D (2012) Compa ison o slope ins abili y sc een-
ing ools ollowing a la ge s o m e en and applica ion o o es man-
agemen and policy. Geomo phology 145–146:115–122. h ps:// doi.
o g/ 10. 1016/j. geomo ph. 2012. 01. 001
Yang H, Yang T, Zhang S e al (2020) Rain all-induced landslides and
deb is lows in Mengdong Town, Yunnan P o ince, China. Landslides
17:931–941. h ps:// doi. o g/ 10. 1007/ s10346- 019- 01336-y
Yin K, Chen L, Zhang G (2008) Regional landslide haza d wa ning and
isk assessmen . Ea h Sci F on 14:85–93. h ps:// doi. o g/ 10. 1016/
s1872- 5791(08) 60005-6
Yu F-C, Chen T-C, Lin M-L e al (2006) Landslides and ain all cha ac e is-
ics analysis in Taipei ci y du ing he yphoon Na i E en . Na Haza ds
37:153–167. h ps:// doi. o g/ 10. 1007/ s11069- 005- 4661-0
Rosa M. Palau (*) · Ma c Be engue · Daniel Sempe e‑To es
Cen e o Applied Resea ch in Hyd ome eo ology, Uni e si a
Poli ècnica de Ca alunya, Jo di Gi ona 1-3 (C4), 08034 Ba celona,
Spain
Email: [email p o ec ed]
Rosa M. Palau · Ma c Be engue · Ma cel Hü limann ·
Daniel Sempe e‑To es
Depa men o Ci il and En i onmen al Enginee ing, Di ision
o Geo echnical Enginee ing and Geosciences, Uni e si a Poli ècnica
de Ca alunya, Jo di Gi ona 1-3 (D2), 08034 Ba celona, Spain
Rosa M. Palau
Na u al Haza ds Di ision, No wegian Geo echnical Ins i u e,
N-0806 Oslo, No way
