267
P edic ing he global dis ibu ion and in asion scena ios o he
Spo ed Lan e n ly, Lyco ma delica ula (Whi e, 1845)
(Hemip e a, Fulgo idae)
En ico Ruzzie 1,2 , Da ide Scaccini3, Pie o Ti ozzi4, Vale io O ioli4, Oli ia Dondina2,4 ,
And ea Di Giulio1,2 , Albe o Pozzebon3, Luciano Bani2,3
1 Depa men o Science, Roma T e Uni e si y, Rome, I aly
2 Na ional Biodi e si y Fu u e Cen e (NBFC), Pale mo, I aly
3 Depa men o Ag onomy, Food, Na u al esou ces, Animals, En i onmen (DAFNAE), Uni e si y o Padua, Pado a, I aly
4 Depa men o Ea h and En i onmen al Sciences, Uni e si y o Milano-Bicocca, Milan, I aly
Co esponding au ho s: En ico Ruzzie (en ico. uzzie @uni oma3.i ); Oli ia Dondina ([email p o ec ed])
Copy igh : © En ico Ruzzie e al.
This is an open access a icle dis ibu ed unde
e ms o he C ea i e Commons A ibu ion
License (A ibu ion 4.0 In e na ional – CC BY 4.0).
Resea ch A icle
Abs ac
Lyco ma delica ula is a species na i e o he PR o China ha has become in asi e in Sou h Ko ea,
Japan, and he Uni ed S a es. I is conside ed a signi ican h ea o ag icul u e, pa icula ly he
i icul u e indus y, and i s u he sp ead in o new a eas could exace ba e i s economic impac .
Species Dis ibu ion Models (SDMs) a e commonly used o analyze he dis ibu ion o in asi e
species, and while L. delica ula has al eady been s udied using his spa ial modeling app oach,
p e ious esea ch has mos ly ocused on he e ec s o bioclima ic a iables, o en o e looking he
ole o habi a cha ac e is ics, including he p esence o hos plan s. In ou s udy, o assess he
p esence o o he sui able habi a s o he species on a global scale, we de eloped a wo-s ep SDM
calib a ed wi hin i s na i e ange, wi hin biogeog aphic ba ie s. In he i s s ep, we buil wo
sepa a e models: one o habi a sui abili y (HSM), inco po a ing land co e , ele a ion, and hos
plan s, and ano he o bioclima ic sui abili y (BSM). The HSM calib a ion included he alloca-
ion o a po ion o pseudo-absences (PAs) along backg ound poin s (BPs) o mi iga e he sampling
bias o occu ences concen a ed in highly u banized a eas (one bias-con olled PA o e e y ou
andom BPs). In con as , he BSM calib a ion was pe o med using a ully andom alloca ion o
BPs. In he second s ep, he wo models we e combined o p oduce an o e all sui abili y map o
he na i e ange, which demons a ed excellen alida ion pe o mance. When p ojec ed globally,
he model con i med ha he species has al eady colonized sui able a eas. Beyond he cu en ly
in aded egions (i.e., No h Ame ica and Eas Asia), he model iden i ies addi ional colonizable
a eas exclusi ely in Eu ope. To e alua e he in asion dynamics in bo h cu en ly in aded egions
and po en ial u u e in asion a eas in Eu ope, we de eloped a esis ance-dis ance cons ained dis-
pe sal model based on a) en i onmen al esis ance, de i ed om he o e all sui abili y map p o-
duced by he SDM, and b) in insic dispe sal dis ance, es ima ed om he empo al p og ession
o occu ences in he p ima y in asion a ea in he eas e n Uni ed S a es. The maximum annual
dispe sal dis ance was es ima ed o be app oxima ely 25 kilome e s. Gi en i s ela i ely low annual
dispe sal capaci y, he species appea s capable o na u ally sp eading wi hin a eas o medium o
high sui abili y, while i s po en ial o expansion in o low-sui abili y a eas emains qui e limi ed.
Medium- and high-sui abili y egions whe e he species is al eady p esen a e likely o unde go
signi ican coloniza ion o e he nex 10 yea s. In con as , ad en i e popula ions in low-sui abili y
a eas seem unable o expand success ully wi hou human-media ed ansloca ions. In he absence
o an adap i e capaci y o h i e in ecological con ex s di e en om i s na i e ange (niche change
o niche shi ), he species appea s o be a ela i ely modes in ade beyond i s cu en ly in aded
Academic edi o : Samuel Wa d
Recei ed:
29 Ma ch 2025
Accep ed:
11 Sep embe 2025
Published:
4 No embe 2025
Ci a ion: Ruzzie E, Scaccini D,
Ti ozzi P, O ioli V, Dondina O, Di
Giulio A, Pozzebon A, Bani L (2025)
P edic ing he global dis ibu ion and
in asion scena ios o he Spo ed
Lan e n ly, Lyco ma delica ula (Whi e,
1845) (Hemip e a, Fulgo idae).
NeoBio a 103: 267–298. h ps://doi.
o g/10.3897/neobio a.103.154246
NeoBio a 103: 267–298 (2025)
DOI: 10.3897/neobio a.103.154246
Ad ancing esea ch on alien species and biological in asions
A pee - e iewed open-access jou nal
NeoBio a
268
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
a eas o Eu ope. Success ully managing he in asion o L. delica ula he e o e depends on he im-
plemen a ion o ea ly de ec ion and e adica ion measu es, which should be deployed as p omp ly
as possible, especially in highly sui able in oduc ion a eas.
Key wo ds: Aucheno hyncha, bioclima ic sui abili y model, habi a sui abili y model, hos
plan s, insec pes , in asi e species, Resis ance-Dis ance Cons ained Dispe sal Model, Species Dis-
ibu ion Model (SDM)
In oduc ion
The Spo ed Lan e n ly, Lyco ma delica ula (Whi e, 1845) (Hemip e a: Fulgo idae)
is a plan hoppe endemic o he PR o China ha has become widely in asi e in Ja-
pan, Sou h Ko ea, and he Uni ed S a es (Lee e al. 2019; U ban and Leach 2023).
The species is o p ima y phy osani a y conce n gi en i s impac on c op p o-
duc ion caused by eeding nymphs and adul s (Lee e al. 2019; U ban and Leach
2023; Zhang e al. 2023). Indeed, bo h nymphs and adul s o L. delica ula eed
on plan sap, a ec ing he heal h o he plan (Da a e al. 2015). Fu he mo e, he
g ega ious beha io o his species, wi h agg ega ions o hund eds o indi iduals,
u he ampli ies he damage o he hos plan (Lee e al. 2019). Lyco ma delica ula
is widely polyphagous, eeding on mul iple plan species, bo h in i s na i e and
in aded ange (CABI 2020; Jung e al. 2022; EPPO 2023), al hough Ailan hus
al issima (Mill.) Swingle (commonly known as T ee o Hea en) is o en e e ed o
as he p ima y o p e e ed hos (Nixon e al. 2022; Elsensohn e al. 2023). As a
consequence, his insec cu en ly occu s in a ious an h opogenic en i onmen s,
such as ag icul u al and u ban a eas, as well as in na u al and semi-na u al habi a s
(U ban and Leach 2023).
Al hough he Spo ed Lan e n ly is po en ially ha m ul o many plan species,
mul iple s udies ca ied ou in he in aded a eas ha e shown ha L. delica ula is
only mode a ely ha m ul o ha dwood and ui ees (Hoo e e al. 2023; Nixon
e al. 2023a) while i is a p ima y pes on g ape ine plan s (Vi is spp.) (Ha ne e
al. 2022). In he eas e n Uni ed S a es, L. delica ula p e e s ineya d-domina ed
landscapes (Ha ne e al. 2022) whe e i signi ican ly con ibu es o yield losses
and g ape ine decline (U ban 2020). These condi ions, he e o e, make his pes
a eal h ea o he wine and g ape indus y on a global scale, bo h in coun ies
whe e i has al eady been in oduced and in hose whe e i could be in he nea
u u e (e.g., Reynolds e al. 2021; Jones e al. 2022; EPPO 2023). Managemen o
his pes p edominan ly elies on inc eased use o b oad-spec um insec icides wi h
high esidual oxici y, posing signi ican h ea s o bene icial non- a ge o ganisms
and he en i onmen (Lee e al. 2019; U ban and Leach 2023). Howe e , his ype
o managemen may be ine ec i e o e he long e m, as ea ed a eas can be e-
colonized by esiden popula ions in adjacen non-c op habi a s (Pa k e al. 2009).
I is c ucial o de e mine he po en ial dis ibu ion o L. delica ula on a global
scale o iden i y he a eas wi h he highes es ablishmen isk and o guide mo e
e ec i e ea ly de ec ion s a egies and pes managemen ac ions (U ban and Leach
2023). In his ega d, species dis ibu ion models (SDMs) ha e become e ec i e
ools o o ecas ing he po en ial dis ibu ion o in asi e insec s (e.g., Eli h 2017).
To da e, se e al s udies p edic ed he dis ibu ion o L. delica ula h ough SDMs,
bo h locally and globally, albei h ough di e en app oaches and da ase s (Jung e
269
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
al. 2017; Namgung e al. 2020; Wakie e al. 2020; Hu on e al. 2022; Cha ois e
al. 2024; Zhao e al. 2024). In addi ion, Jones e al. (2022) and Mon gome y e
al. (2023), beyond assessing species sui abili y, also es ima ed he isk o in asion
by accoun ing o connec i i y unde cu en and u u e scena ios. These s udies,
in some cases, es ima ed sui abili y based on bioclima ic a iables, while in o he s,
hey employed a modelling amewo k ha ended o o e emphasize he in luence
o hese a iables o e habi a ea u es (e.g., ele a ion, hos plan s). As a esul , he
in e p e a ion o L. delica ula po en ial dis ibu ion is o en con ined o a b oad
scenopoe ic con ex a he han a local scale. Recen s udies highligh ha inco po-
a ing habi a a iables in o a wo-s ep amewo k, sepa a ing a habi a sui abili y
model (HSM) om a bioclima ic sui abili y model (BSM), enhances SDM o e-
cas ing (Adde e al. 2023; Ruzzie e al. 2024). Fo phy ophagous species, in eg a -
ing hos plan s in o habi a modelling imp o es he eliabili y o p edic ions o
po en ially sui able a eas, bo h wi hin and ou side he na i e ange o he a ge
species (Dang e al. 2021; Ruzzie e al. 2024, 2025).
S udies analyzing he mo emen s o adul emales in sea ch o o iposi ion si es
(Wol in e al. 2019) o ood esou ces (Wol in e al. 2020) highligh he low p open-
si y o L. delica ula o co e long dis ances. Howe e , he dispe sal po en ial o p op-
agules emains unknown. Gaining insigh in o he spa ial expansion dynamics o
p opagules could suppo he de elopmen o e ec i e con ainmen s a egies, pa -
icula ly in ela ion o al eady in es ed a eas and he sui abili y o adjacen habi a s.
The aims o his esea ch we e:
1) o sepa a ely calib a e a BSM and a HSM wi hin he na i e ange o L. delica-
ula as a basis o p oduce an o e all en i onmen al sui abili y map;
2) o o ecas sui able a eas o he species a global scale based on he models
(BSM and HSM) calib a ed o he na i e ange, wi h a pa icula ocus on
in aded and newly colonized a eas;
3) o e alua e he model e ec i eness in p edic ing he cu en dis ibu ion o
his species in in aded a eas (USA and Sou h Ko ea);
4) o es ima e he annual dispe sal dis ance o L. delica ula based on eposi o y
occu ences;
5) o model he in asion ange expansion and o ecas he u u e USA, Sou h
Ko ea and Japan in asion scena ios, as well as he po en ial in asion scena io
in Eu ope.
Ma e ials and me hods
Species occu ences
Occu ences o L. delica ula in i s na i e and in aded anges we e e ie ed om
he Global Biodi e si y In o ma ion Facili y (GBIF) (h ps://doi.o g/10.15468/
dl.azzdjh; 32,752 occu ences) and iNa u alis (41,867 occu ences), accoun ing
o 74,619 occu ences wo ldwide. The occu ence da a used in he SDMs we e
upda ed o Oc obe 11, 2024. The da ase was checked o all po en ial issues
(e.g., duplica e eco ds, misiden i ica ions, e c.). Fo he BSM calib a ion on na-
i e ange (mainland PR o China), we used all a ailable occu ences, while o he
HSM, we used a subse o he p e ious da a, selec ing only hose wi h coo dina es
in deg ees wi h a leas wo decimal places o esolu ion o accu acy lowe han
270
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
2.5 km (hal o he as e cell size o he en i onmen al laye s used in he HSM
model) (Ruzzie e al. 2025). The a ional o he p ocedu e elies on he capabili y
o associa ing each occu ence wi h habi a da a de i ed om land co e ca og a-
phy wi h a g id esolu ion o 0.04167° (co esponding o app oxima ely 5 km a
he equa o ). Fo BSM, occu ences wi h highe unce ain y we e also included,
as bioclima ic a iables end o exhibi g ea e spa ial ine ia in hei a iabili y.
The o e all da ase was hinned h ough he unc ion “ensemble.spa ialThin” o
he package “Biodi e si yR” (Kind and Coe 2005), wi h a hinning pa ame e o
2.5 km in R, e sion 4.4.2 (R Co e Team 2023) o exclude duplica ed da a and
emo e occu ences wi hin he same as e cell. Fo HSM and BSM 324 and 402
occu ences we e used, espec i ely.
Occu ences om Japan, he Ko ean peninsula, and Taiwan we e excluded om
he analyses since hese a eas a e ou side he pu a i e na i e ange o his species; in
hese coun ies, L. delica ula is na u ally absen , because o he exis ence o majo
biogeog aphic ba ie s (ocean and moun ain anges) ha he species is unable o
o e come, whe eas p esen occu ences a e he esul o human in oduc ion.
Bioclima ic, ele a ion and habi a a iables
Bioclima ic (19 s a a) and ele a ion (i.e., digi al ele a ion model, DEM) laye s
we e downloaded om Wo ldClim 2.1 (a ailable a : h p://www.wo ldclim.com/
e sion2) and a e aged o he yea s 1970–2000, wi h a g id esolu ion o 0.00833°.
Since mos Bioclima ic s a a a e s ongly collinea , a p eselec ion o bioclima ic
a iables may help in educing mul icollinea i y, imp o ing model in e p e abili y,
and enhancing p edic i e accu acy (Do mann e al. 2013; Plisco e al. 2014).
We p eselec ed a subse o 10 among hem: mean diu nal ange (mean o mon h-
ly (max empe a u e – min empe a u e)) [BIO2]; max empe a u e o wa mes
mon h [BIO5]; min empe a u e o coldes mon h [BIO6]; mean empe a u e o
wa mes qua e [BIO10]; mean empe a u e o coldes qua e [BIO11]; annual
p ecipi a ion [BIO12]; p ecipi a ion o we es qua e [BIO16]; p ecipi a ion o
d ies qua e [BIO17]; p ecipi a ion o wa mes qua e [BIO18]; p ecipi a ion
o coldes qua e [BIO19].
Habi a a iables we e e ie ed om The Global 1-km Consensus Land Co e
a ailable a h p://www.ea hen .o g. Da a we e ep esen ed by 12 laye s (e e -
g een/deciduous needlelea ees [EDNT], e e g een b oadlea ees [EBT], decid-
uous b oadlea ees [DBT], mixed/o he ees [MOT], sh ubs [SHR], he baceous
ege a ion [HERB], cul i a ed and managed ege a ion [CMV], egula ly looded
ege a ion [RFV], u ban/buil -up [BU], ba en lands [BR], open wa e s [OW]
and Snow/Ice [SI]), each o which p o ides he p e alence o one land-co e class,
exp essed as a pe cen age in a pixel ha ing a esolu ion o 30 a c-seconds (i.e.,
0.00833° esolu ion).
To educe compu a ional e o , bioclima ic, DEM, and land co e laye s we e
upscaled o a 0.04167° esolu ion ia he “agg ega e” unc ion included in he e a
package (Hijmans 2024). Among he 12 land-use laye s, we excluded a p io i SI
and OW om he HSM modelling, as hey ep esen unsui able en i onmen s o
he a ge species and o insec s in gene al.
To habi a a iables, we added a laye o hos plan s o es whe he his co a i-
a e could imp o e he p edic i e e ec i eness o HSM (see Ruzzie e al. 2024,
2025). To his aim, we compiled he hos plan s lis o L. delica ula (see Suppl.
271
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
ma e ial 1: able S1) using in o ma ion ob ained om he Eu opean and Medi-
e anean Plan P o ec ion O ganiza ion (EPPO 2015), whose da ase is based on
scien i ic li e a u e, expe assessmen s, and o icial epo s p o ided by Na ional
Plan P o ec ion O ganiza ions o i s membe coun ies. I mo e han one species
belonging o he same genus was known as a hos plan , o i congene ic species
could be used as ica ian hos plan species in a eas a isk o in asion o al eady
in aded, he genus was used as he hos axon (e.g., Ace ). O he wise, a single
species was used as he hos plan (e.g., A. al issima). Fo hos plan s iden i ied
a he genus le el, all eco ds o each species wi hin he genus we e downloaded
om GBIF using he “occ_da a” unc ion in he gbi R package (Chambe lain
e al. 2024). This same app oach was applied o hos plan s a he species le el.
Da a o each hos plan axon we e combined in o a single laye o geo e e enced
poin s using he “as.ppp” unc ion o c ea e a densi y map ia ke nel in e pola ion
wi h he “densi y.ppp” unc ion; he ke nel in e pola ion adius was es ima ed wi h
he “bw.ppl”; all unc ions pe aining o spa s a R package (Baddeley and Tu ne
2005; Baddeley e al. 2005). To educe po en ial o e sampling, we used he loga-
i hmic ans o ma ion o he ke nel densi y maps ( ep esen ing he quan i a i e
ange map o each hos plan axon). To p e en o e es ima ion o each hos axon
dis ibu ion, we mul iplied he loga i hmic densi y laye by he land co e ype
associa ed wi h each axon habi (e.g., deciduous b oadlea ees) o by he land
co e ype ele an o economically aluable axa (e.g., cul i a ed and managed
ege a ion); he associa ion be ween hos plan axa and land co e is indica ed in
Suppl. ma e ial 1: able S1. Finally, we p oduced an o e all hos plan s laye by
ex ac ing he maximum densi y alue ound in each g id cell ac oss all hos plan
laye s. Since hos plan s can locally eplace each o he , we used he maximum
densi y alue a he han summing he densi y alues o all laye s o a oid o e es-
ima ing hos plan densi y. Da a on hos plan axa we e downloaded om GBIF
on Sep embe 30, 2024. I should be acknowledged ha hos plan occu ences
a e shaped by bioclima ic condi ions, and ha he ke nel densi y map implici ly
inco po a es hese e ec s. As a esul , bioclima ic in luences on he insec species
may be indi ec ly accoun ed o in he HSM.
To a oid collinea i y issues, all en i onmen al co a ia es used o build he HSM
and BSM we e subjec ed o a iance in la ion ac o (VIF) analysis using he “ i-
co ” unc ion om he usdm R package (Naimi e al. 2014). Highly co ela ed
a iables, hose wi h a Pea son’s co ela ion coe icien | | > 0.7, we e excluded
(Do mann e al. 2013). This analysis was conduc ed sepa a ely wi hin he calib a-
ion ex en s o BSM (b oad- ange ex en : 85°E o 125°E; 0°N o 70°N) and HSM
(na ow- ange o na i e ex en : 90°E o 125°E; 25°N o 45°N).
None o he land use co a ia es used in he SDMs we e pai wise co ela ed.
Following he VIF analysis, six ou o en bioclima ic a iables we e e ained o
he BSM p ocedu e: h ee ela ed o empe a u e (BIO2, BIO5, and BIO11) and
h ee ela ed o p ecipi a ion (BIO17, BIO18, and BIO19).
Species Dis ibu ion Modelling amewo k
The modelling app oach was based on an ensemble p ocedu e (A aújo and New
2007; Guisan e al. 2017). To be e assess he in luence o en i onmen al a iables
on L. delica ula spa ial dis ibu ion, bioclima ic and habi a a iables (including
digi al ele a ion model) we e modeled sepa a ely in wo dis inc ensemble models
272
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
(HSM and BSM), and hei esul ing maps combined as a spa ial p oduc , ollow-
ing he amewo k ou lined by Ruzzie e al. (2024). We included DEM among
he habi a a iables o accoun o local mic oclima ic a ia ions. This app oach is
based on he idea ha in eg a ing wide-scale (bioclima ic) and na ow-scale (habi-
a ) a iables in o a single model may esul in unde es ima ing he con ibu ion o
a iables ac ing a di e en geog aphic scales (scenopoe ic s local; sensu Ho al e
al. 2010, see F anklin 2010). F om a p obabilis ic pe spec i e, his s a egy elies on
he assump ion ha bioclima ic sui abili y, i.e., he p obabili y o inding an occu -
ence a a gi en si e acco ding o he BSM [P(BSM)], ep esen s a necessa y bu no
su icien condi ion (a clima ic cons ain ), while habi a sui abili y, as es ima ed
by he HSM and in insically condi ioned by bioclima ic ac o s [P(HSM|BSM)],
e ines he local p obabili y. Acco dingly, he condi ional p obabili y is ob ained by
mul iplying he HSM and BSM p obabili y (sui abili y) maps [P(BSM|HSM)=P(B-
SM)•P(HSM|BSM)] (see G imme and S i zake 2001). We acknowledge ha he
o e all sui abili y map, ob ained as he p oduc o he BSM and HSM p obabili y
maps, does no ep esen a o mal condi ional p obabili y bu a he a composi e
sui abili y index, since he wo models we e de eloped independen ly. Ne e he-
less, because HSM is in insically condi ioned by bioclima ic ac o s, he composi e
sui abili y index can be conside ed concep ually close o a condi ional p obabili y.
When modelling p esence-only da a, he selec ion o pseudo-absences (PAs)
o backg ound poin s (BPs) is one o he mos c i ical aspec s o add ess (Ba -
be -Massin e al. 2012); pseudo-absences should be in ended as loca ions whe e
he species is assumed o be absen , hough i was ne e con i med h ough ob-
se a ion, while backg ound poin s ep esen andomly sampled loca ions ac oss
he s udy a ea whe e he p esence o absence o he species is unknown (Guisan e
al. 2017). Sampling PAs o BPs exclusi ely wi hin a species’ na i e ange p e en s
a p ope e alua ion o he in luence o bioclima ic a iables, which p ima ily op-
e a e a b oad spa ial scales. As a esul , he inabili y o isola e he e ec s o hese
bioclima ic ac o s would hinde he iden i ica ion o key a iables ha limi he
species’ na i e ange and educe he accu acy o p edic ing po en ially sui able
a eas beyond ha ange. The e o e, when wo king a la ge spa ial scales (e.g., con-
inen al o global), i is essen ial o sample PAs o BPs wi hin he occu ences ange
and sligh ly beyond i s limi s in o de o de e mine whe he , and o wha ex en ,
bioclima ic a iables cons ain he species’ dis ibu ion (see Guisan e al. 2017;
Adde e al. 2023). In con as , a iables such as land co e , DEM, and hos plan s
exhibi ine-scale a ia ion and ac p ima ily a he local le el, in luencing habi a
sui abili y. To assess he e ec s o hese local-scale co a ia es, PAs o BPs should be
selec ed as close as possible o known p esence poin s (Fou nie e al. 2017; Ma eo
e al. 2019; Adde e al. 2023). This app oach minimizes he po en ial con ounding
e ec o bioclima ic a iables on he es ima ion o habi a -le el in luences.
Since we empi ically obse ed a ela i ely high densi y o occu ences o L. deli-
ca ula in u ban a eas, we suspec ed a possible sampling bias likely due o an easie
de ec ion o his species in u ban con ex han in mo e emo e a eas (e.g., Beck
e al. 2014). To e alua e his obse a ion, we conduc ed an ini ial modelling a -
emp using BPs. As expec ed, his i s ou pu e ealed a subs an ial o e es ima-
ion o a iables associa ed wi h an h opogenic con ex s, pa icula ly BU ( esul s
no shown). To o e come his issue, we decided o use a mixed se o BPs and PAs
o HSM calib a ion, i.e. a andom BPs iden i ica ion combined wi h PAs iden-
i ied on he base o a sampling bias map. We i s iden i ied an HSM calib a ion
273
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
a ea (wi hin he na ow- ange ex en ; Suppl. ma e ial 2: ig. S1) de ined by su -
aces whe e he densi y o occu ences exceeded 0.001 poin s/km2. Subsequen ly,
we de e mined a se o PAs, in which he numbe o PAs equaled he numbe o
occu ences. The alloca ion o PAs was p opo ional o he ac ional co e o BU
wi hin each cell o he calib a ion a ea. Addi ionally, we iden i ied a se o BPs,
whe e he numbe o BPs equaled ou imes he numbe o occu ences wi hin he
calib a ion a ea. O e all, he o al numbe o PAs + BPs o HSM was i e imes
he numbe o occu ences (324 PAs + 1296 BPs). Fo he BSM, we iden i ied
BPs ac oss he wide- ange ex en , se ing he o al numbe o BPs a i e imes he
numbe o occu ences (2010 BPs; Suppl. ma e ial 2: ig. S2).
Bo h HSM and BSM models e u n a map o sui abili y in e ms o species
p esence p obabili y. F om he spa ial p oduc o he wo maps o species p esence
p obabili ies, we ob ained an o e all sui abili y map, which e u ns he p obabili y
o inding he species in each loca ion based on bo h habi a and bioclima ic co-
a ia es (see Ruzzie e al. 2024 o de ails).
The modelling p ocedu e was pe o med ia he biomod2 package e sion 4.2-
5-2 (Thuille e al. 2024). As modelling algo i hms, he A i icial Neu al Ne -
wo k (ANN), he Flexible Disc iminan Analysis (FDA), he Gene alized Addi i e
Model (GAM), he Gene alized Boos ing Model (GBM), Random Fo es (RF),
he Mul iple Adap i e Reg ession Splines (MARS), and he Maximum En opy
(MAXNET) we e used, all unde “bigboss” se ings (Thuille e al. 2024). A o-
al o 20 se s o PAs + BPs we e used o HSM, and 5 se s o BPs we e used o
BSM. Gi en he lack o ex e nal da a o e alua e he models’ accu acy, a 5- old
c oss- alida ion da a-spli ing p ocedu e was pe o med using 80% o he da a
as he aining se wi h he emaining 20% as he alida ion se (Ruzzie e al.
2024). Each o he models ob ained om bo h HSM and BSM was alida ed ia
wo alida ion me ics: he a ea unde he ecei e ope a ing cha ac e is ic (ROC)
cu e (AUC; Hanley and McNeil 1982) and he ue skill s a is ic (TSS, Allouche
e al. 2006). F om he wo me ics, we selec ed he bes 10% models in alida ion
o p oduce he inal ensemble models, one o each me ic (Allouche e al. 2006;
Mand eka 2010). Habi a and bioclima ic ensemble models we e calcula ed ia
he “EMwmean” me hod in biomod2 package. To accoun o he pe o mance o
each single model used in building he ensemble model, we weigh ed i s con ibu-
ion by applying a ‘p opo ional’ op ion decay. A h ee- un pe mu a ion p ocedu e
was used o assess he impo ance o each a iable included in he ensemble model.
The same ensemble model amewo k was adop ed o bo h he HSM and BSM.
A o al o 1,400 models we e de eloped (7 modelling algo i hms × 20 se s o
(PAs + BPs) × 5 c oss- alida ion uns × 2 alida ion me ics) o he HSM, and
350 models (7 modelling algo i hms × 5 se s o BPs × 5 c oss- alida ion uns × 2
alida ion me ics) o he BSM.
The HSM and BSM we e p ojec ed on o he ollowing i e geog aphic ex en s:
1) na i e a ea; 2) global, 3) Eas e n and Sou heas Asia (ESEA), 4) he Uni ed S a es
o Ame ica (USA), and 5) Eu ope. The ESEA p ojec ion p o ides a de ailed map o
he na i e ange and he closely in aded a ea (Japan, Sou h Ko ea and Taiwan). The
USA p ojec ion p o ides a map o a eas al eady in aded by L. delica ula and hose
sui able o species es ablishmen ; we op ed o p ojec he calib a ed model o e
he Eu opean e i o y since his species seems o be p esen in Spain (iNa u alis
eco d) and because i s po en ial impac on majo Eu opean ag icul u al c ops,
pa icula ly g ape ine (Vi is spp.) (e.g., Hu on e al. 2022; U ban and Leach 2023).
274
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
Based on all o he geog aphic ex en s, we p oduced he o e all sui abili y map.
Since he ou pu o bo h HSM and BSM p oduced sui abili y maps exp essed
in e ms o p obabili y o occu ence o L. delica ula s a ing om independen
da ase s, he o e all sui abili y maps we e ob ained as he p oduc o he HSM
and BSM maps. Using he Boyce index (Boyce e al. 2002), we assessed he e ec-
i eness o he o e all sui abili y map in p edic ing he dis ibu ion o L. delica ula
wi hin he calib a ion a ea. We also applied he index o e alua e he p edic i e
pe o mance o he o e all sui abili y maps ob ained by p ojec ing he calib a ed
models on o in aded egions, including he eas e n Uni ed S a es, whe e a ma-
jo sel -sus aining popula ion is es ablished, and Sou h Ko ea. The Boyce index
was used in an explo a o y manne o he in asi e popula ions, as hese a e s ill
expanding and many a eas iden i ied as sui able by he model ha e no ye been
colonized. This index measu es a model’s abili y o disc imina e among a eas o
a ying sui abili y by compa ing p edic ed alues a p esence loca ions wi h hose
expec ed unde a andom dis ibu ion. Analyses we e pe o med using he modE A
package (Ba bosa e al. 2013). To de ec possible ex apola ion e ec s o he HSM
and BSM a he global scale, we used he mul i a ia e en i onmen al simila i y
su ace index (MESS; Eli h e al. 2010), implemen ed in he dismo package (Hi-
jmans e al. 2024).
Resis ance-Dis ance Cons ained dispe sal models
To e alua e he in asion model o L. delica ula in he USA and Sou h Ko ea,
whe e mul iple es ablished popula ions a e p esen , we de eloped a amewo k o
assess he impac o en i onmen al esis ance on he dispe sal capaci y du ing he
coloniza ion p ocess. This app oach was no applied in Japan due o he ecen
es ablishmen o his pes in ha egion.
The amewo k elies on h ee s eps: 1) he es ima ion o he annual dispe sal
dis ance; 2) he iden i ica ion o he bes unc ion de ining he esis ance su ace
wi h espec o species coloniza ion (i.e., species ange expansion); 3) he eachable
h eshold, de ined as cumula i e cos aced by p opagules du ing he coloniza ion
dispe sal ( his aspec accoun s o bo h he ecological pe meabili y o he en i on-
men al ma ix and he in insic dispe sal dis ance o L. delica ula). The in asion
model was hen used in o ecas ing po en ial in asion scena ios in in aded o po-
en ially in adable a eas ac oss he wo ld.
Dispe sal dis ance
F om a managemen pe spec i e, besides es ima ing he annual ange expansion,
i could also be ele an o assess he magni ude o in insic linea dispe sal o
L. delica ula. To his end, om he o e all global da ase we used occu ence da a
om he eas e n USA (11,326 eco ds), he only egion wo ldwide whe e he in-
asion p ocess is well documen ed, o es ima e he annual ange expansion and
co esponding expansion a e. These da a a e consis en wi h hose epo ed by De
Bona e al. (2023).
The annual dispe sal dis ance was es ima ed by analyzing da a om wo consec-
u i e yea s, unde he assump ion ha occu ences we e de ec ed immedia ely ol-
lowing local popula ion es ablishmen . Fo each poin in he da ase co esponding
o yea i, he closes poin in he da ase om yea i+1 was iden i ied, and he Eu-
275
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
clidean dis ance be ween hem was calcula ed. S a ing om 2015 (when mo e han
one occu ence was a ailable), we selec ed and bonded among consecu i e yea s all
he closes pai wise Euclidean dis ances. Then, we boo s apped he o e all median
o hese Euclidean dis ances, along wi h i s con idence in e als a 95% (E on and
Tibshi ani 1993). In addi ion, we boo s apped he 99 h pe cen ile o he closes
pai wise Euclidean dis ances, and i s con idence in e als a 95%. This p ocedu e al-
lowed us o es ima e he maximum dispe sal dis ance o L. delica ula om one yea
o ano he . All poin s om he yea i+1, whe e he alue o he closes pai wise Eu-
clidean dis ances exceeded he 99 h pe cen ile, we e classi ied as an e en o new in-
oduc ion due o human-media ed ansloca ion (see “Resis ance cos h eshold”).
Resis ance su aces
Resis ance su aces a e widely used o assess how much en i onmen al ea u es
hinde animal dispe sal, o bo h assess he popula ion ecological connec i i y
(Bani e al. 2015; Lecis e al. 2022, 2024) and o e alua e he coloniza ion p o-
cesses (She pa e al. 2020). To achie e his, we calcula ed a se o esis ance maps
as unc ion o he o e all sui abili y map ob ained om he SDM amewo k. We
es ed wo me hods o e alua e he esis ance encoun e ed by p opagules du ing
dispe sal: a) esis ance de ined as he complemen a y alue o sui abili y (i.e., ad-
di i e in e se ela ionship; Zelle e al. 2012; Bani e al. 2018), and b) esis ance
ha a ies non-linea ly wi h sui abili y, ollowing a logis ic unc ion (McRae and
Beie 2007; Zelle e al. 2012):
a) R = 1−S
b) R = 1 / (1 + exp(−k * (1−S−x0)))
whe e R is he esis ance, S is he sui abili y, k is he s eepness and x0 is he in lec-
ion poin ha de e mines a change o conca i y. In his case we used h ee se s o
alues wi h di e en alues o x0 (0.25, 0.5, 0.75), while we se k alue equal o 15
(Fig. 1). The a ionale behind his concep is ha coloniza ion begins by p io i iz-
ing a eas wi h he lowes esis ance, hose wi h he highes sui abili y (i.e., highes
p obabili y o coloniza ion), e en i hey a e geog aphically a he away bu s ill
wi hin he dispe sal dis ance achie able by p opagules. Coloniza ion hen ex ends
o egions wi h lowe o subop imal sui abili y only a e hese high-sui abili y
a eas ha e been sa u a ed (Lo ell e al. 2021). In addi ion, we assumed ha newly
colonized cells immedia ely ac as new sou ces o p opagule sp ead.
We hen de e mined he alue o esis ance be ween wo neighbo ing cells o he
esis ance map, p oducing a ansi ion ma ix, h ough he unc ion “ ansi ion”
o he gdis ance R package ( an E en 2017); in he ansi ion ma ix, o a pai o
neighbo ing cells A and B he ansi ion alue be ween hem (i.e., he ecological
esis ance expe ienced by an o ganism passing om he cell A o cell B) depends
on he esis ance alues o he wo cells and he ansi ion unc ion de ining he
ma hema ical ela ionship be ween he wo cell alues; as he ansi ion unc ion
we used he ecip ocal (i.e., he in e se ela ionship; Ad iaensen e al. 2003; McRae
and Beie 2007) o he a e age alue o he wo adjacen cells: (x)=1/mean x, whe e
x is a ec o con aining he alues o wo neighbo ing cells. The spa ial dis ance
be ween wo cells is also conside ed: o a 3×3 g oup o cells, he dis ance be ween
282
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
cos h eshold. F om jus unde 2,000 km2 colonized in 2015, he in aded a ea in
2024 had expanded app oxima ely 70- old, eaching nea ly 130,000 km2. These
a eas include only occu ences associa ed wi h p opagules p esumed o o igina e
om he ini ial in oduc ion (i.e., he a ea enclosed by he isocho es, as illus a -
ed in Fig. 8).
Table 1. Values o a ibu es ob ained om he unc ions used in assessing he expansion model o Lyco ma delica ula in he eas e n USA.
In i alics a e shown he bes h esholds selec ed by he sco e o each unc ion and ela ed pa ame e s used in de ining he esis ance su ace.
In bold i alics: he bes h eshold associa ed wi h he bes unc ion.
Func ions Resis ance h esholds and spa ial a ibu es Values o he esis ance h esholds and de i ed spa ial a ibu es
Linea Resis ance cos h eshold 10,500 11,000 11,500 12,000 12,500 13,000
In aded ange (km2) 108,418 116,888 125,280 134,093 141,776 149,750
Poin included in o he ange (%) [A] 78.2 82.3 86.4 89.3 91.2 92.9
Poin s/ ange (poin s/100km2) [B] 6.47 6.32 6.19 5.98 5.77 5.57
Sco e [A×B] 506.5 520.3 534.1 533.7 526.4 517.2
Logis ic: Resis ance cos h eshold 17,000 17,250 17,500 17,750 18,000 18,250
k = 15; x0 = 0.25 In aded ange (km2) 133,350 136,314 139,363 142,570 145,342 148,463
Poin included in o he ange (%) [A] 85.5 86.5 87.9 88.7 89.3 89.9
Poin s/ ange (poin s/100km2) [B] 5.77 5.70 5.66 5.58 5.51 5.44
Sco e [A×B] 492.2 492.9 497.4 494.8 492.4 488.8
Logis ic: Resis ance cos h eshold 1,100 1,150 1,200 1,250 1,300 1,350
k = 15; x0 = 0.50 In aded ange (km2) 95,782 98,538 101,283 101,283 101,283 108,114
Poin included in o he ange (%) [A] 89.6 91.6 93.1 93.9 94.4 94.8
Poin s/ ange (poin s/100km2) [B] 5.74 5.87 5.97 6.02 6.05 6.08
Sco e [A×B] 537.1 546.0 548.4 545.4 540.0 532.8
Logis ic: Resis ance cos h eshold 25 30 35 40 45 50
k = 15; x0 = 0.75 In aded ange (km2) 98,368 117,700 140,666 163,061 181,944 201,574
Poin included in o he ange (%) [A] 68.8 83.8 91.5 93.1 93.4 93.6
Poin s/ ange (poin s/100km2) [B] 6.27 6.39 5.83 5.12 4.61 4.17
Sco e [A×B] 431.5 534.8 533.6 477.0 430.6 389.8
Table 2. A ea in aded by L. delica ula in eas e n USA om 2015 o 2024 s a ing om i s ini ial
in oduc ion poin . The ange a ea was es ima ed by he ange expansion model and e e s solely o
he yea -by-yea p og ession o he popula ions de i ed om p opagules whose o igin is pu a i ely
a ibu ed o he ini ial in oduc ion (i does no include a eas occupied by popula ions belie ed o
ha e o igina ed a e in e nal ansloca ions media ed by humans).
Yea In aded a ea (km2)
2015 1,797
2016 8,611
2017 21,041
2018 35,724
2019 52,363
2020 72,511
2021 89,480
2022 102,769
2023 115,525
2024 127,721
283
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
Figu e 8. Range expansion model: yea -by-yea 2017 o 2024 in asion dynamic o L. delica ula in eas e n USA s a ing om i s in o-
duc ion. The isocho e lines de i ed om he ange expansion model e e solely o he popula ion de i ed om p opagules whose o igin is
pu a i ely a ibu ed o he ini ial in oduc ion (i does no include a eas occupied by popula ions belie ed o ha e o igina ed om in e nal
ansloca ions media ed by humans). Red poin s ep esen he occu ences ha a e pu a i ely assigned o he popula ion o igina ing om
he i s in oduc ion e en , based on he maximum annual dispe sal dis ance es ima ed o L. delica ula (~25 km).
Fu u e in asion scena ios in he USA, Sou h Ko ea and Japan, and
po en ial in asion scena io in Eu ope
The expansion scena ios o L. delica ula o e he nex 10 yea s in he USA, Sou h
Ko ea and Japan, assessed by he ange expansion model and s a ing om i s
cu en dis ibu ion, a e shown in Figs 9–11, espec i ely. The Eu opean scena io,
based on h ee hypo he ical en y poin s and one iNa u alis eco ding poin o
L. delica ula (Mad id, Spain), and de eloped using he eas e n USA ange expan-
sion model, is illus a ed in Fig. 12. This in asion scena io is based on he assump-
ion ha popula ion expansion equi es p io local es ablishmen . Based on he
eas e n USA ange expansion model, he po en ial sp ead o L. delica ula in he
Ibe ian Peninsula is unlikely due o he low sui abili y o he in oduc ion a ea.
284
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
Figu e 9. Fu u e in asion scena io o Lyco ma delica ula in he USA and Sou he n Canada o e he nex 10 yea s, based on he species’
cu en dis ibu ion.
Figu e 10. Fu u e in asion scena io o Lyco ma delica ula in Sou h Ko ea o e he nex 10 yea s, based on he species’ cu en dis ibu ion.
All hese scena ios should be conside ed wi hou accoun ing o po en ial hu-
man-media ed ansloca ions, which could u he acili a e he expansion o
L. delica ula ange.
285
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
Figu e 11. Fu u e in asion scena io o Lyco ma delica ula in Japan o e he nex 10 yea s, based on he species’ cu en dis ibu ion.
Figu e 12. Possible Eu opean in asion scena io o Lyco ma delica ula, assuming h ee hypo he ical en y poin s 10 yea s a e local popu-
la ion es ablishmen om Milano Malpensa Ai po (I aly), Pa is Cha les de Gaulle Ai po (F ance), and Hambu g Po (Ge many). Red
poin s indica e pu a i e occu ences o L. delica ula.
286
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
Discussion
Species dis ibu ion model: compa a i e e alua ion
Al hough he ou pu o HSM buil solely on habi a a iables may be pa ially in-
luenced by bioclima e (hos plan axa and land co e s can be a ec ed by clima ic
condi ions o some ex en ), bioclima ic a iables used o build he independen
BSM can none heless p o ide a sepa a e con ibu ion o he dis ibu ion o L. del-
ica ula, dis inc om ha o e ed by habi a a iables.
Lyco ma delica ula seems o be p ima ily in luenced by bioclima ic condi ions
in i s na i e ange. Howe e , wi hin a eas ha a e bioclima ically sui able, habi a
a ailabili y appea s o be a key limi ing ac o in he species dis ibu ion, as indi-
ca ed by he habi a sui abili y map.
When conside ing he o e all sui abili y map, he combined impac o bo h
bioclima ic and habi a ac o s on he species dis ibu ion pa e n becomes e en
mo e e iden . Consequen ly, i seems unlikely ha he species could na u ally ex-
pand whe e all i s ecological needs a e no simul aneously me , as can be obse ed
by he o e all sui abili y map wi hin na i e ange. In addi ion, wi hin i s na i e
ange, L. delica ula seems unable o expand in o nea by a eas wi h sui able habi-
a s and bioclima es when ecological ba ie s, such as oceans o majo moun ain
anges, a e p esen .
Buil -up a eas (BU) eme ged as an impo an a iable in he calib a ion a ea
(i.e., he na i e ange), showing a posi i e e ec on he p obabili y o species p es-
ence. This e ec is also con i med by o he au ho s, who ha e obse ed a simila
in luence in a eas subjec o in asion (Owen e al. 2024; Le ine e al. 2025). Ac-
co ding o hese s udies, L. delica ula may bene i om he u ban en i onmen ,
pa icula ly due o he p esence o hos plan s used o o namen al pu poses, as
well as he mic oclima ic “hea island” e ec , which could inc ease he chances o
su i al o he species. In addi ion, he mean empe a u e o he coldes qua e
(BIO11) signi ican ly in luences he win e su i al o L. delica ula, as i s eggs
equi e a pe iod o o e win e ing o b eak diapause. Howe e , p olonged and ex-
eme cold e en s can lead o emb yo mo ali y, as p e iously obse ed by se e al
au ho s (Lee e al. 2011, 2014; Keena and Nielsen 2021; K ei man e al. 2021).
The nume ous non-na i e a eas wi h simila bioclima ic and habi a condi ions
iden i ied as sui able by ou modelling amewo k can hus be colonized by he
species only h ough human-media ed ansloca ion, due o geog aphic sepa a ion
(Ladin e al. 2023).
The adop ed app oach appea s o e ec i ely iden i y he mos sui able a eas o
L. delica ula on a global scale, as all cu en ly in aded a eas a e classi ied as sui able
o , a leas , mode a ely sui able bo h in e ms o bioclima ic ac o s and habi a , in-
cluding hos plan s (o e all p obabili y o species p esence ≥ 0.5). Fo his eason,
i can be hypo hesized ha L. delica ula may no be su icien ly plas ic o change
o expand i s ecological niche.
Ou esul s, like se e al ecen s udies (Namgung e al. 2020; Wakie e al. 2020;
Hu on e al. 2022; Cha ois e al. 2024; Zhao e al. 2024; Lin and Liao 2025),
di e om Jung e al. (2017), who linked Lyco ma delica ula mainly o opical
and sub opical humid clima es. In con as , mos s udies, including ou s, associ-
a e he species wi h a b oade ange om humid con inen al o humid sub opical
clima es. These di e ences likely s em om Jung e al. use o he CLIMEX model,
while o he s employed algo i hms like Maxen , Random Fo es and o he s.
287
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
Ou s udy o e s impo an ad ancemen s compa ed o p e ious analyses in
modeling he dis ibu ion o L. delica ula. Unlike Namgung e al. (2020), who
in eg a ed bo h bioclima ic and habi a a iables (limi ed o Vi is ini e a L. as he
hos plan ) wi hin a single model ocused solely on Sou h Ko ea, ou app oach
models hese ac o s sepa a ely, allowing o a mo e de ailed and di e en ia ed as-
sessmen . Fu he mo e, we u ilize a composi e laye comp ising se e al dozen hos
plan axa, in con as o he single hos conside ed by Namgung e al., and apply
an ensemble modeling app oach, enhancing p edic ion obus ness. Finally, he al-
loca ion o PAs o BPs likely in luenced he disc epancies obse ed be ween he
wo s udies, as i is well known ha modelling echniques based solely on p esence
da a a e highly a ec ed by his aspec (as no ed by S een e al. 2024).
Wakie e al. (2020) de eloped a bioclima ic model inco po a ing p esence da a
om bo h he na i e ange and in aded egions, wi h backg ound poin s an-
domly dis ibu ed wi hin a la ge polygon su ounding he occu ence da a. Thei
esul s a e gene ally consis en wi h ou BSM, excep o he na i e ange, whe e
hey p edic no ably lowe sui abili y. Wakie e al. acknowledge ha hea y eliance
on non-na i e da a in hei model likely con ibu es o he unde es ima ion o
sui abili y in he na i e ange.
Ou BSM map la gely ag ees wi h ha o Hu on e al. (2022), which was p o-
duced by a e aging wo di e en maps: he i s based on bioclima ic a iables,
digi al ele a ion model, o es canopy heigh , and access o ci ies, and he second
ela ed o he sui abili y o Ailan hus al issima as a hos species. The alignmen
be ween ou BSM map and he o e all sui abili y map by Hu on e al. (2022) is
la gely due o he dominan in luence o he bioclima ic co a ia es in he model
ha p oduced he la e map p o ided by Hu on and co-au ho s.
The sui abili y map p oduced by Zhao e al. (2024), based on a Maxen model
in eg a ing 19 bioclima ic a iables and a digi al ele a ion model, and using oc-
cu ence da a om bo h na i e and in aded egions, la gely aligns wi h ou BSM
map in e ms o global sui abili y pa e ns. Howe e , hei map gene ally shows
lowe sui abili y alues compa ed o ou s. A key limi a ion o hei app oach is he
andom global alloca ion o backg ound poin s, which isks placing hese poin s
in clima ically sui able a eas ha emain uncolonized due o biogeog aphic ba i-
e s o he ecen absence o in oduc ions. This leads o an unde es ima ion o he
impo ance o a iables in luencing ue habi a sui abili y (see Ruzzie e al. 2025;
S een e al. 2024). Ne e heless, since mos sui able egions a e al eady pa ially o
ully in aded by L. delica ula, his issue likely had minimal impac on hei o e all
esul s, as mos backg ound poin s ell wi hin unsui able a eas. Ou app oach,
howe e , e ec i ely minimizes his ca ea .
Cha ois e al. (2024) used an ensemble model based on bioclima ic a iables
o p oduce a global sui abili y map la gely consis en wi h ou BSM and Zhao
e al. (2024). Howe e , while Zhao e al.’s map shows gene ally lowe sui abili y
han ou s, Cha ois e al.’s map indica es highe sui abili y in bo h in ensi y and
geog aphic ex en . Thei s udy p ope ly alloca ed PAs o bioclima ic modeling bu
did no exclude egions sepa a ed by biogeog aphic ba ie s, which may be sui able
bu uncolonized. The key di e ence wi h ou BSM likely s ems om hei use o
PAs es ic ed o he no h and sou h o he p esence polygon along a la i udinal
g adien , whe eas we alloca ed backg ound poin s (BPs) ac oss he en i e g adien .
This comp ehensi e BP alloca ion in ou model enables a mo e balanced assess-
men o bioclima ic e ec s, while hei me hod may o e emphasize condi ions
288
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
wi hin he p esence polygon by assuming i is en i ely sui able and a eas ou side
a e unsui able.
The model we de eloped did no iden i y he U.S. s a e o Cali o nia as a c i ical
a ea o po en ial in asion isk, despi e some de ec ions o L. delica ula epo ed in
li e a u e and da abases. We belie e hese occu ences likely ep esen ad en i ious
popula ions, as hey ha e no shown any signs o local ange expansion. Such oc-
cu ences may ha e been sus ained by comme cial anspo a ion (see Jones e al.
2022; Ladin e al. 2023; Mon gome y e al. 2023) wi hin a egion o low sui abil-
i y. Con e sely, o he s udies published a ound he same ime (Wakie e al. 2020;
Jones e al. 2022; Mon gome y e al. 2023) iden i ied Cali o nia as a sui able a ea
o his pes , based on modelling app oaches ha included occu ences om pop-
ula ions no ye a equilib ium. Howe e , we a gue ha calib a ing models wi h
occu ences om non-equilib ium popula ions, such as ad en i e popula ions,
may dis o niche es ima ion and he in luence o co a ia es, po en ially leading
o misleading p edic ions. Simila ly, p opagules de ec ed in cen al U.S. s a es and
Flo ida a e in a eas o low sui abili y and hus ep esen a low in asion isk. The
same pa e n applies o Canada, whe e con i med occu ences occu mos ly in
a eas o low sui abili y, excep o No a Sco ia, he wes e n sho es o Lake On a io,
and he no hwes e n sho es o Lake E ie.
Ul ima ely, Namgung e al. (2020), Hu on e al. (2022), Jones e al. (2022),
Mon gome y e al. (2023), and Cha ois e al. (2024) should be c edi ed o
acknowledging he impo ance o inco po a ing o assessing hos plan s, albei
h ough di e en app oaches, when e alua ing he sui abili y o po en ial coloni-
za ion by L. delica ula.
The analy ical amewo k we ha e de eloped is, in ou iew, capable o p o id-
ing a mo e obus p edic i e capaci y, hanks o a se ies o me hodological aspec s,
many o which ep esen no el implemen a ions. In pa icula , he calib a ion o
he model using da a om popula ions in equilib ium wi hin he na i e ange o
he species allows o a mo e ealis ic assessmen o he ecological niche occupied
by he species. Fu he mo e, he c i e ia adop ed o iden i y BPs and PAs, along
wi h hei spa ial ex en and ela ionship o occu ence eco ds, play a c ucial ole
in e ining he es ima ion o he in luence o en i onmen al co a ia es. The in-
eg a ion o hos plan dis ibu ions, combined wi h high- esolu ion land co e
da a, u he enhances he es ima ion o habi a -le el sui abili y, p o iding a mo e
accu a e ep esen a ion o he ecological equi emen s o he species. Addi ionally,
he sepa a e implemen a ion o a HSM and a BSM, wi hin a wo-s ep modelling
app oach, enables a mo e e ec i e dis inc ion be ween he e ec s o local-scale
co a ia es (add essed in he HSM) and hose associa ed wi h b oade -scale ecolog-
ical p ocesses (add essed in he BSM). Al oge he , his modelling s a egy leads o
an imp o ed explana o y unde s anding o he dis ibu ion o L. delica ula wi hin
he na i e ange and deli e s mo e e ec i e p edic i e pe o mance, in line wi h a
niche conse a ion pe spec i e, in egions ou side he na i e a ea.
E alua ion o SDM p ojec ions
The Boyce index, employed o assess he e ec i eness o he SDM calib a ed on
he na i e ange in p edic ing he dis ibu ion o L. delica ula in in aded egions
(USA and Sou h Ko ea), e eals wo dis inc scena ios. In he USA, al hough his
insec has ye o sa u a e sui able habi a s (wi h popula ions s ill expanding), he
289
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
e alua ion is highly posi i e, as a signi ican p opo ion o occu ences all wi hin
a eas iden i ied as sui able. In con as , he Boyce index o Sou h Ko ea indica es
ha he dis ibu ion o occu ences is a om sa u a ing sui able a eas and does
no exhibi a pa e n consis en wi h na u al popula ion expansion. Ins ead, he
occu ences appea highly sca e ed, likely esul ing om mul iple in oduc ion
e en s and/o human-media ed ansloca ions, which ha e ini ia ed o a e sus-
aining a ious in asion ho spo s o may ep esen ad en i e popula ions. Gi en
he ecen in asion o Japan by L. delica ula, e alua ing he accu acy o he model
p edic ion o his egion no longe seems app op ia e.
Despi e he p esence o se e al sui able a eas on a global scale, many o hese a -
eas, which a e cu en ly no in aded, appea o be po en ially suscep ible o coloni-
za ion, in ag eemen wi h Hu on e al. (2022). Se e al o hese a eas a e cha ac e -
ized by he p esence o aluable c ops on which his pes could cause a signi ican
economic damage (e.g., g ape ine). These a eas ha e a mild clima e, wi h empe a-
u es highe han −11 °C du ing win e ( he minimum su i al h eshold o he
insec , as indica ed by Lee e al. 2014), and a e also human-domina ed, wi h an
ex ensi e p esence o non-na i e plan s, bo h in asi e and o namen al, used as hos
plan s by L. delica ula (Nixon e al. 2023a). These condi ions hus g ea ly suppo
L. delica ula popula ions, c ea ing sui able habi a s and acili a ing he expansion
o his pes by o ming in asion co ido s, in ag eemen wi h Zhang e al. (2023).
Conside ing he in asion p ocess o L. delica ula in ESEA, Sou h Ko ea is al-
eady a la gely in aded e i o y whe e i can be conside ed nea ly ubiqui ous, e en
a p esen , wi h popula ions anging om ad en i e o es ablished. In Taiwan, he
isk o in asion is a he limi ed due o he low sui abili y ac oss mos o he e -
i o y, excep o he an h opized con ex o he no he n pa o he island (e.g.,
Taipei and he su ounding a eas), whe e ad en i e popula ions al eady exis . In
con as , o Japan, conside ing he high sui abili y o much o he e i o y and
he p esence o es ablished popula ions, he isk o in asion appea s signi ican . All
hese po en ial isks o in asion should be in e p e ed in he ligh o he cu en
clima ic and habi a condi ions.
Range expansion model and in asion scena ios
Conside ing bo h he sui abili y and he in insic dispe sal po en ial o L. delica -
ula, we simula ed he insec ’s ange expansion in in aded a eas, s a ing om he
cu en dis ibu ion, as well as in po en ially in adable a eas, o e a 10-yea pe iod.
Ou p ojec ions o he Uni ed S a es, whe e in asi e popula ions a e mo e
widesp ead, iden i y wo majo egions a isk o coloniza ion. The i s is loca ed
eas o he Appalachian Moun ains, ex ending owa d he A lan ic coas , and ap-
pea s o be al eady ex ensi ely colonized, wi h L. delica ula ha ing nea ly occupied
all sui able habi a s. The second lies u he inland, wes o he Appalachians and
sou h o he G ea Lakes, whe e he in asion is ongoing, showing clea di ec ion-
al mo emen owa d he con inen al in e io . Smalle , likely in e nally eloca ed
popula ions a e al eady p esen he e, and he model p edic s nea ly comple e col-
oniza ion o he mos sui able a eas wi hin he nex 10 yea s. The model also indi-
ca es a low likelihood o popula ions in Cali o nia, he Cen al-Wes e n s a es, and
Flo ida de eloping in o expanding in asi e popula ions.
In Sou h Ko ea, he 10-yea expansion scena io p ojec ed om cu en ly colo-
nized a eas sugges s ha po en ial in asion will likely concen a e along he sou h-
290
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
wes e n coas . O he popula ions, possibly a ising om in e nal ansloca ions,
a e unlikely o con ibu e o u he ange expansion, including in he Seoul a ea,
despi e high occu ence densi ies eco ded he e. Gi en hese unce ain ies, his
in asion pa e n should be in e p e ed wi h cau ion, bu we deemed i aluable o
p esen his po en ial scena io in ela ion o habi a sui abili y ac oss Sou h Ko ea.
Fo Japan, whe e sui abili y is high, he 10-yea scena io indica es ha cu -
en popula ions could apidly sp ead h ough he cen al and sou he n egions o
Honshu Island. Howe e , he model does no p edic ull sa u a ion o sui able a -
eas u he no h wi hin his pe iod, so addi ional ange expansion is s ill possible.
In Eu ope, he ange expansion simula ion emphasizes ha success ul in asion
depends on es ablished popula ions wi hin medium o highly sui able a eas. This
is illus a ed by he single eco ded occu ence in Mad id, Spain, which ou model
sugges s is unlikely o con ibu e o u he sp ead. Con e sely, in oduc ions and
es ablishmen s in highly sui able a eas could lead o apid expansion, as demon-
s a ed by he h ee hypo he ical en y poin s included in ou simula ions.
Pes s a us pe spec i es
The indings o his s udy sugges ha , gi en he limi ed ecological plas ici y and
ela i ely slow expansion a e es ima ed o L. delica ula, he p omp implemen a-
ion o e adica ion measu es could be pa icula ly e ec i e in elimina ing ad en-
i e popula ions. These e o s mus be ein o ced by s ingen measu es o p e en
he es ablishmen o new in asion ho spo s caused by acciden al human-media -
ed dispe sal. Addi ionally, in mode a ely o highly sui able a eas, u he e o s
a e equi ed, including he implemen a ion o e ec i e moni o ing p o ocols and
e adica ion me hods as soon as he species is de ec ed.
The damage caused by L. delica ula in asion should be assessed conside ing i s
po en ial impac on c ops and o namen al plan s, de i ed om i s eeding ac i i y
and honeydew deposi ion, especially when i agg ega es in huge numbe s (U ban
and Leach 2023).
As highligh ed by he sui abili y model and he in asion scena ios, a subs an ial
pa o he USA emains exposed o he isk o in asion wi h po en ially simila
impac s o hose obse ed in o he in aded a eas, as indica ed by Ha pe e al.
(2019), Leach and Leach (2020), U ban (2020), U ban and Leach (2023), and
Mol ini e al. (2024). Global ade inc eases he likelihood o L. delica ula in o-
duc ions and associa ed economic losses in coun ies wi h ex ensi e sui able a eas
o his species, pa icula ly key p oduce s o ag icul u al and i icul u al p oduc s,
as emphasized by Hu on e al. (2022). This scena io appea s especially ealis ic o
Eu ope, gi en he ecen eco d o his pes in Spain.
Ou indings also align wi h b oade in asion ecology p inciples and wi h e i-
dence om o he in asi e he bi o es. In many species, bioclima ic condi ions de-
e mine he b oad limi s o po en ial sp ead, while habi a and hos a ailabili y ac
as ine -scale il e s. Fo example, he sp ead o he Spongy Mo h, Lyman ia dispa
(Linnaeus, 1758), ac oss No h Ame ica depends no only on clima ic sui abili y
bu also on he dis ibu ion o hos ee species (Régniè e e al. 2012), and he Em-
e ald Ash Bo e , Ag ilus planipennis Fai mai e, 1888, emains igh ly cons ained by
he a ailabili y o F axinus hos s despi e sui able clima es (He ms and McCullough
2014). Simila ly, ba k bee le, Dend oc onus ponde osae (Hopkins, 1902), dynamics
291
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
(illus a e how clima e de ines physiological bounda ies, while local o es composi-
ion go e ns ou b eak se e i y (Ben z e al. 2010). These pa allels unde line ha in-
asion isk o L. delica ula should no be e alua ed on clima ic sui abili y alone, bu
on he combined il e ing ac ion o clima e, habi a , and dispe sal oppo uni ies.
Clima ic condi ions se he physiological limi s o su i al and de elopmen , while
he p esence and densi y o sui able hos plan s a ec success ul coloniza ion and
ep oduc i e ou pu . Local habi a ea u es and composi ion u he egula e popu-
la ion g ow h and expansion, ul ima ely shaping he speed and pa e n o in asion.
Recognizing he in e play o hese ac o s is he e o e essen ial o ully unde s and
he mechanisms unde lying L. delica ula sp ead and o accu a ely assess in asion
isk. Ou in eg a ed HSM–BSM amewo k e lec s his mul i-laye ed il e ing
p ocess, emphasizing ha clima ically sui able a eas may no be colonized unless
app op ia e hos s and habi a s a e p esen , and ha na u al sp ead emains limi ed
by majo geog aphic ba ie s unless o e come by human-media ed ansloca ion.
As p edic ed by ou scena io, he in oduc ion o L. delica ula in o mul iple
sui able a eas ac oss he Eu opean subcon inen could igge a apid pes expan-
sion, possibly esul ing in signi ican impac s on c op p oduc ion. Coun ies like
F ance, I aly, and Ge many, among he wo ld’s la ges wine p oduce s, a e pa ic-
ula ly ulne able o he in asion isk o L. delica ula, as a signi ican po ion o
hei e i o y is highly sui able o his pes . The acciden al in oduc ion o L. del-
ica ula in hese egions could indeed cause se e e damage o he wine p oduc ion
sec o , leading o subs an ial losses as documen ed in Pennsyl ania, USA (Leach
and Leach 2020), whe e, in 2018, up o 90% o g apes we e los despi e he use
o insec icides (U ban 2020). The po en ial impac on he ui p oduc ion sec o
should also no be unde es ima ed. Howe e , hese impac s may depend on he
ea u es o he ag oecosys em in ques ion, including managemen s a egies, as
discussed in Nixon e al. (2023b) and U ban and Leach (2023).
Conclusions
The SDM and he sp ead model implemen ed in his s udy ep esen aluable
ools o managing L. delica ula and in asi e species mo e b oadly, as hey enable
he p ecise iden i ica ion o a eas a highes isk o in asion, including bo h coloni-
za ion and expansion phases. This would suppo he de elopmen and implemen-
a ion o ea ly su eillance and a ge ed moni o ing s a egies, op imizing esou ce
alloca ion and enhancing he e ec i eness o in e en ions, while educing un-
necessa y cos s in low- isk a eas. Mo eo e , mapping en i onmen al sui abili y is
c ucial o e ec i ely planning po en ial in oduc ions o biological con ol agen s,
by iden i ying a eas whe e such ac ions a e mos likely o succeed. The ou pu s
o hese models also p o ide a solid basis o suppo poli ical and managemen
decisions, o e ing use ul insigh s o de ining qua an ine zones, egula ing he
mo emen o goods, o launching ou each and awa eness campaigns o local
communi ies. Looking ahead, in eg a ing hese models wi h socio-economic da a,
such as anspo a ion lows o comme cial ac i i ies, could u he imp o e p e-
dic i e accu acy by mo e p ecisely iden i ying human-media ed dispe sal ou es
and po en ial en y poin s. Finally, ex ending hese app oaches o u u e clima e
scena ios would allow an icipa ion o in asion dynamics o e he medium o long
e m, enabling p oac i e and adap i e planning o managemen s a egies.
298
NeoBio a 103: 267–298 (2025), DOI: 10.3897/neobio a.103.154246
En ico Ruzzie e al.: Global in asion scena ios o Spo ed Lan e n ly
Supplemen a y ma e ial 1
Supplemen a y ables
Au ho s: En ico Ruzzie , Da ide Scaccini, Pie o Ti ozzi, Vale io O ioli, Oli ia Dondina, And ea Di
Giulio, Albe o Pozzebon, Luciano Bani
Da a ype: xlsx
Explana ion no e: able S1. Schema ic associa ions be ween he hos plan axon and land co e .
able S2. E alua ion HSM. Table o he h ee e alua ion me ics used o HSM. able S3. E al-
ua ion BSM. Table o he h ee e alua ion me ics used o BSM. able S4. Impo ance o he
co a ia es included in he HSM. able S5. Impo ance o he co a ia es included in he BSM.
Copy igh no ice: This da ase is made a ailable unde he Open Da abase License (h p://openda a-
commons.o g/licenses/odbl/1.0/). The Open Da abase License (ODbL) is a license ag eemen
in ended o allow use s o eely sha e, modi y, and use his Da ase while main aining his same
eedom o o he s, p o ided ha he o iginal sou ce and au ho (s) a e c edi ed.
Link: h ps://doi.o g/10.3897/neobio a.103.154246.suppl1
Supplemen a y ma e ial 2
Supplemen a y igu es
Au ho s: En ico Ruzzie , Da ide Scaccini, Pie o Ti ozzi, Vale io O ioli, Oli ia Dondina, And ea Di
Giulio, Albe o Pozzebon, Luciano Bani
Da a ype: pd
Explana ion no e: ig. S1. Dis ibu ion o he 20 se s o PAs + BPs used o he habi a sui abili y
model (HSM) calib a ion in he na ow na i e calib a ion ange. ig. S2. Dis ibu ion o he 5 se s
o BPs used o he bioclima ic sui abili y model (BSM) calib a ion in he wide calib a ion ange.
ig. S3. Response cu es o co a ia es included in he Lyco ma delica ula habi a sui abili y models
(HSM). ig. S4. Response cu es o co a ia es included in he Lyco ma delica ula bioclima ic
sui abili y models (BSM). ig. S5. Habi a sui abili y map o Lyco ma delica ula in he na ow
na i e calib a ion ange ob ained p ojec ing he habi a sui abili y model (HSM). ig. S6. Biocli-
ma ic sui abili y map o Lyco ma delica ula in he wide calib a ion ange ob ained p ojec ing he
bioclima ic sui abili y model (BSM). ig. S7. O e all sui abili y map o Lyco ma delica ula in he
na ow na i e calib a ion ange. ig. S8. Boyce o he o e all sui abili y map o Lyco ma delica-
ula in he na ow na i e calib a ion ange. ig. S9. Global (a) HSM and (b) BSM Mul i a ia e
En i onmen al Simila i y Su aces (MESS) index. ig. S10. O e all sui abili y map o Lyco ma
delica ula in he Eas e n USA. In ed species occu ences. ig. S11. Boyce o he o e all sui abili y
map o Lyco ma delica ula in he Eas e n USA. ig. S12. O e all sui abili y map o Lyco ma del-
ica ula in Sou h Ko ea. In ed species occu ences. ig. S13. Boyce o he o e all sui abili y map
o Lyco ma delica ula in Sou h Ko ea. ig. S14. Wo ld p ojec ions o Lyco ma delica ula sui abili y
maps; a) habi a sui abili y map; b) bioclima ic sui abili y map. ig. S15. Eas e n and sou heas
Asia (ESA) p ojec ions o Lyco ma delica ula sui abili y maps. ig. S16. USA p ojec ions o Ly-
co ma delica ula sui abili y maps. ig. S17. De ails o USA (Eas Coas ) p ojec ions o Lyco ma
delica ula sui abili y maps. ig. S18. Eu opean p ojec ions o Lyco ma delica ula sui abili y maps.
Copy igh no ice: This da ase is made a ailable unde he Open Da abase License (h p://openda a-
commons.o g/licenses/odbl/1.0/). The Open Da abase License (ODbL) is a license ag eemen
in ended o allow use s o eely sha e, modi y, and use his Da ase while main aining his same
eedom o o he s, p o ided ha he o iginal sou ce and au ho (s) a e c edi ed.
Link: h ps://doi.o g/10.3897/neobio a.103.154246.suppl2
