Deli e able 5.5
Scien i ic documen summa ising he esul s o
he Species dis ibu ion modelling and mapping
o he alien species impac s on he Eu opean
coas al and ma ine habi a s
Ve sion 1.0
Da e o deli e y
2025-06-30
Au ho s names
Ma hias Obs 1
A ilia ions
1 Depa men o Ma ine Sciences, Uni e si y o Go henbu g, Box 461, 405 30 Go henbu g, Sweden;
PUBLIC DOCUMENT
Re . A es(2025)7308469 - 05/09/2025
Documen In o ma ion
G an Ag eemen
101082021
P ojec Ac onym
MARCO-BOLO
P ojec Ti le
MARine COas al BiOdi e si y Long- e m Obse a ions
Deli e able Numbe
D5.5
Wo k Package Numbe
WP5
Deli e able Ti le
Scien i ic documen summa ising he esul s o he Species dis ibu ion
modelling and mapping o he alien species impac s on he Eu opean coas al
and ma ine habi a s
Lead Bene icia y
Uni e si y o Go henbu g, Pa ne Numbe 9
Au ho (s)
Ma hias Obs (Uni e si y o Go henbu g)
Due Da e
30.05.2025
Submission Da e
30.06.2025
Dissemina ion Le el
PU
Type o Deli e able
R
Ve sion 0.5
05.05.2025, Ma hias Obs
Ve sion 0.6
25.06.2025, Ma hias Obs , Ca lo a Muñiz
Ve sion 1.0
30.06.2025, Ma hias Obs
Repo Edi
3
Execu i e Summa y
This deli e able summa izes he scien i ic ou pu s gene a ed by ask T5.4.“Mapping he impac s o
Non-Indigenous species (NIS) on Eu opean coas al and ma ine habi a s, based on mul i-disciplina y
app oaches”. In his ask we modelled 82 species iden i ied by ma ine gene ic moni o ing p og ams
in 2020-2024 as occu ing ou side hei na i e ange. Models p oduced p edic ions o sui able
habi a in cu en and u u e clima e scena ios o hese non-indigenous species (NIS) and can be
used o iden i y egions wi h ele a ed isk o es ablishmen and sp ead o alien species.
Repo Edi
4
Con en s
Documen In o ma ion ............................................................................................................................ 2
Execu i e Summa y ................................................................................................................................. 3
1. Objec i e .......................................................................................................................................... 5
2. Backg ound ...................................................................................................................................... 5
3. Me hodology ................................................................................................................................... 5
3.1. Modelling app oach ...................................................................................................................... 5
4. Resul s and Discussion ..................................................................................................................... 7
4.1. Modelling ial ............................................................................................................................... 7
4.2. Sho alls o he cu en app oach ............................................................................................... 8
4.3. Fu u e wo k................................................................................................................................... 8
4.4. Conclusion ..................................................................................................................................... 8
5. Resea ch ou pu s ............................................................................................................................. 9
5.1. Summa y able .............................................................................................................................. 9
5.2. Con e ence con ibu ions ............................................................................................................. 9
6. Appendix ........................................................................................................................................10
6.1. Table 1.........................................................................................................................................10
7. Re e ences .....................................................................................................................................12
Repo Edi
5
1. Objec i e
The objec i e o his ask was o un a new analysis wi h he in asi e alien species (IAS) ho spo model
o a numbe o non-indigenous species (NIS) ecen ly de ec ed by gene ic moni o ing campaigns and
p oduce maps showing a eas o po en ial in asion and ecological impac o hese NIS in Eu opean
coas al wa e s.
2. Backg ound
Ma ine ecosys ems a e a ec ed by human ac i i ies in he sea such as ishing, aquacul u e and
shipping. These p ocesses o en cause in asions o alien species (also called non-indigenous species
o NIS), which al e na i e communi ies and lead o he global decline o biodi e si y. Common
app oaches o p e en and mi iga e ma ine in asions include isk assessmen s and ea ly wa ning
p og ams o alien species. Recen ly, se e al isk assessmen ools o alien species ha e been
de eloped based on species dis ibu ion modelling me hods (SDMs). He e s a is ical models a e used
o analyse he po en ial habi a sui abili y o alien species and he eby p edic po en ial ange
ex ensions and a eas o po en ial impac . He e we model he po en ial dis ibu ion o 80 non-
indigenous species de ec ed by he gene ic moni o ing p og ammes, ARMS-MBON and he Swedish
na ional po moni o ing p og amme (Pagnie e al 2025; Sundbe g e al 2024) and hei abili y o
o ecas po en ial ange expansions o hese species in Eu opean coas al wa e s.
3. Me hodology
3.1. Modelling app oach
Modelling o ho spo s was ca ied ou using Species Dis ibu ion Models (SDM). To his end we
de eloped a wo k low o iden i y po en ial high- isk a eas o he sp ead and es ablishmen o alien
species in Eu opean coas al wa e s. The SDM wo k low is based on a modelling app oach de eloped
by he Uni e si y o Go henbu g (Ka lsson e al 2019; Leidenbe ge e al 2015) and was u he
adap ed o deal wi h p ope ies o alien species dynamics, o example iden i y op imal alien species
moni o ing si es (Be gk is e al 2020) and p edic ansi ions be ween eshwa e and b ackish and
ma ine habi a s (Obs & Ande sson 2023). The modelling esul s no only show po en ial dis ibu ion
a eas o indi idual alien species bu can also be used o map egions whe e sui able habi a s o
known alien species o e lap. Such egions wi h a high o e all in asion isk, i.e. high isk o in oduc ion
and es ablishmen , can be conside ed in asi e ho spo s.
The models a e based on a Random Fo es machine lea ning algo i hm. A de ailed explana ion o all
analy ical s eps included in he modeling can be ound in Leidenbe ge e al (2015) and Ka lsson e al
(2019).
Species-speci ic models we e c ea ed based on each species' maximum dis ibu ion wi hin he s udy
a ea and he en i onmen al a iables lis ed below. Random Fo es models we e un wi h 10.000
backg ound poin s (including inpu poin s) d awn om he s udy a ea. All models we e se o p oduce
a p obabili y a he han bina y p ojec ion in as e o ma . The models we e es ed using con usion
ma ices and ROC cu es, and hen p ojec ed wi h he same en i onmen al a iables back in o he
Repo Edi
6
s udy a ea. The esul s we e isualized as maps o po en ial dis ibu ion a eas, i.e. geog aphical
p ojec ions o sui able habi a s.
Geog aphical scope
The s udy a ea, i.e. he geog aphic a ea whe e models we e ained o , a e coas al wa e s o Eu ope
including he Black Sea, he Medi e anean Sea, he No h-Eas A lan ic and he Bal ic Sea. P edic ions
o he models we e gene a ed o he same egion.
Documen a ion
The model's sou ce code, as well as inpu s and esul s wi h species-speci ic and g oup-based
p ojec ions, a e a ailable on Gi Hub
h ps://gi hub.com/biomobs /IAS_ho spo _model/ ee/mas e /Modelling_ ial_2025.
Biological da a
The lis o a ge species (Table 1) was ob ained om gene ic moni o ing p og ams ARMS-MBON
(Pagnie e al 2025) p og am and he Swedish na ional po moni o ing campaign (Sundbe g e al
2024).
En i onmen al da a
G idded en i onmen al da a a ailable as global ma ine da a laye s h ough he Bio-O acle websi e
(h p://www.bio-o acle.ugen .be/) wi h a esolu ion o 5 a c-min (Tybe ghein e al 2012) we e used
o gene a e en i onmen al a iables o he SDMs. These da a laye s a e gene a ed om mon hly
sa elli e da a (Aqua-MODIS and SeaWiFS websi e h ps://oceancolo .gs c.nasa.go /) as well as in-si u
measu ed oceanog aphic da a om he Wo ld Ocean Da abase 2009 (Boye e al 2009). The ollowing
da a laye s we e used o cu en and u u e clima e scena io SP119:
Da a laye
Desc ip ion
nos_mean_dep hsu
Mean su ace ni a e in mmol pe m3
chl_mean_dep hsu
Mean su ace chlo ophyll in mg pe m3
02_max_dep hsu
Maximum su ace dissol ed molecula oxygen in mmol pe m3
po4_mean_dep hsu
Mean su ace phospha e in mmol pe m3
he ao_min_dep hsu
Minimum su ace sea wa e empe a u e in ºC
pa _mean_mean_dep hsu
Mean su ace pho osyn he ic a ailable adia ion in Eins ein pe m2 and
day
02_mean_dep hsu
Mean su ace dissol ed molecula oxygen in mmol pe m3
phyc_mean_dep hsu
Mean phy oplank on concen a ion a he su ace
he ao_mean_dep hsu
Mean su ace sea wa e empe a u e in ºC
he ao_max_dep hsu
Maximum su ace sea wa e empe a u e in ºC
ph_mean_dep hsu
Mean su ace PH
02_mindep hsu
Minimum su ace dissol ed molecula oxygen in mmol pe m3
si_mean_dep hsu
Mean su ace silica e in mmol pe m3
02_max_dep hmax
Maximum ben hic dissol ed molecula oxygen in mmol pe m3
02_mean_dep hmax
Mean ben hic dissol ed molecula oxygen in mmol pe m3
02_min_dep hmax
Minimum ben hic dissol ed molecula oxygen in mmol pe m3
phyc_mean_dep hmean
Mean phy oplank on concen a ion a mean dep h.
he ao_mean_dep hmean
Mean midwa e sea wa e empe a u e in ºC
so_mean_dep hmean
Mean salini y a mean dep h
02_mean_dep hmean
Mean ben hic dissol ed molecula oxygen in mmol pe m3
po4_mean_dep hmean
Mean ben hic phospha e in mmol pe m3
Repo Edi
7
so_mean_dep hsu
Mean salini y a he su ace
sws_mean_dep hsu
Mean sea wa e speed a he su ace
siconc_mean_dep hsu
Mean ben hic sea ice co e in ac ion
si hick_
Sea ice hickness in m
siconc_max_dep hsu
Maximum ben hic sea ice co e in ac ion
4. Resul s and Discussion
4.1. Modelling ial
Summa y o esul s
The esul s o he modelling ial a e documen ed on he Gi Hub p ojec page
h ps://gi hub.com/biomobs /IAS_ho spo _model/ ee/mas e /Modelling_ ial_2025 in he olde
“ esul s”
h ps://gi hub.com/biomobs /IAS_ho spo _model/ ee/mas e /Modelling_ ial_2025/ esul s.
Indi idual p ojec ions
Species speci ic p ojec ions can be used o iden i y a eas o sui able habi a whe e species ha e no
ye been obse ed o a eas whe e species may mig a e o in he nea u u e, o example, he Bal ic
coas o Bugula ne i ina (Fig. 1). The maps o p ojec ed sui able habi a o all he species can be
ound in he “Plo s_species” olde :
h ps://gi hub.com/biomobs /IAS_ho spo _model/ ee/mas e /Modelling_ ial_2025/ esul s/Plo s_
species.
Figu e 1. Example o model p edic ions o Bugula ne i ina. Le : cu en ly known dis ibu ion indica ed by ed
do s, wi h pseudo-absence poin s indica ed by blue do s. Colou scale is linea om 0.00 (unsui able habi a )
o 1.00 (highly sui able habi a ). Righ : Di e ence in sui able habi a be ween SSP119 scena io p edic ion o
2100 and cu en clima e scena io. SSP119 e e s o a clima e scena io wi hin he Sha ed Socioeconomic
Repo Edi
8
Pa hways (SSPs) amewo k, speci ically SSP1-RCP1.9, which is a low-emission scena io. Colou scale is linea
om -1.00 (maximum loss o sui able habi a ) o +1.00 (maximum gain o sui able habi a ).
4.2. Sho alls o he cu en app oach
The models a e sensi i e o many o pa ame e s, which means ha he ou pu s p esen ed om his
ial ha e o be in e p e ed wi h cau ion. When applied in decision making, hese models should be
buil om a consensus o se e al algo i hms (e.g., no only Random o es ) and clima e scena ios (e.g.,
no only SPP119) o be con iden abou he p edic ions o sui able habi a s and mig a ion ends.
Visualisa ion o esul s is challenging as he e a e many model ou pu s ha need o be inspec ed
simul aneously o in e p e a ion, such as e.g., es s a is ics, pa ame e se ings, maps in a ious
esolu ions (high, low), o ma s (png, geo i ), and p ojec ion se ings (p esen , u u e scena ios). A
isualisa ion in e ace (e.g., R-shiny) may be app op ia e o sol e his p oblem. Bu such applica ions
should be de eloped only a e he wo k low is consolida ed and end use s o he models ha e
p o ided de ailed speci ica ions o how and when hey would use such an applica ion.
The analy ical wo k low applied he e should be be e documen ed o allow ep oducibili y and
epea abili y and he eby enable swi and ad-hoc p edic ions o ele an species as soon as hese
a e de ec ed in he gene ic moni o ing p og ams. This wo k is cu en ly scheduled o he emainde
o he p ojec .
4.3. Fu u e wo k
The modelling ials need o be be e documen ed and in e p e ed ac oss di e en clima e scena ios.
Also, he indi idual p ojec ions will ha e o be combined o p oduce upda ed maps o he ho spo s.
These a e planned o be used in decision making by Swedish au ho i ies, e.g., when g an ing
exemp ions om ballas wa e ea men in Swedish po s. A wo kshop is planned o Sep embe 2025
o speci y he end use demands o ou pu maps.
Fu he mo e, esul s om he modelling ial will be p esen ed a he 12 h In e na ional Con e ence
on Ma ine Bioin asions in Oc obe 2025.
We cu en ly plan one mo e scien i ic manusc ip which will ocus on he publica ion o he SDM
wo k low an esul s.
Finally, he SDM analysis ool will be added o he EDITO pla o m (h ps://www.edi o.eu) du ing
sp ing 2026 and he eby become an analy ical ool in he Digi al Twin o he Ocean (DTO).
4.4. Conclusion
The gene ic moni o ing p og ams ha dona ed he alien species lis s o his expe imen do no ye
p o ide enough species occu ence da a o in luence he ou come o he model p edic ions, which
means ha GBIF and OBIS s ill p o ide mos o he occu ence da a o he modelling. Howe e , he
Repo Edi
9
gene ic moni o ing p og ams do p o ide he lis o species de ec ed o be “on he mo e”. Fo hese
species he models p o ide aluable knowledge on po en ial alien species impac as hey a e able o
align he p esen ly known dis ibu ion wi h he po en ial dis ibu ion o hese species in he u u e.
5. Resea ch ou pu s
5.1. Summa y able
Ou pu desc ip ion
Re e ence
Documen a ion
Species dis ibu ion
model (SDM)
wo k low o p edic
po en ial high- isk
a eas o
in oduc ion,
es ablishmen and
sp ead o in asi e
species
Be gk is e al (2020); Obs & Ande sson
(2023)
h ps://gi hub.com/biomobs /IAS_ho spo _m
odel/
Sc ip -based
wo k low o iden i y
known alien species
om gene ic
obse a ion da a se s
using WRiMS
Da aghmeh e al (2024) A long- e m
ecological esea ch da a se om he
ma ine gene ic moni o ing p og amme
ARMS-MBON 2018-2020. Mol Ecol Res. In
Re iew.
h ps://gi hub.com/ liz-be-opsci/lw-iji-
in asi e-checke
Da a pape
p esen ing he i s
ba ch o gene ic
obse a o y da a
om EMO BON
Da aghmeh e al (2024) A long- e m
ecological esea ch da a se om he
ma ine gene ic moni o ing p og amme
ARMS-MBON 2018-2020. Mol Ecol Res. In
Re iew.
Molecula Ecology Resou ces 25 (4), e14073,
h ps://doi.o g/10.1111/1755-0998.14073
Da a pape
p esen ing he non-
indigenous species
de ec ed by he ARMS
MBON
Pagnie e al (2025) Using he long- e m
gene ic moni o ing ne wo k ARMS-MBON
o de ec ma ine non-indigenous species
along he Eu opean coas s.
Biological In asions 27, 77 (2025),
h ps://doi.o g/10.1007/s10530-024-03503-2
Con e ence alk a
12 h ICMB, Madei a
Island, Po ugal
Pagnie e al (2025) Assessing he
e ec i eness o gene ic obse a o y
ne wo ks in de ec ing and moni o ing
ma ine non-indigenous species
h ps://ma inebioin asions.in o/p og am
Con e ence alk a
12 h ICMB, Madei a
Island, Po ugal
Obs e al (2025) A science-policy
in e ace o non-indigenous species
moni o ing and managemen .
h ps://ma inebioin asions.in o/p og am
5.2. Con e ence con ibu ions
12 h In e na ional Con e ence on Ma ine Bioin asions, Madei a Island, Po ugal; Da es: 7-9 Oc obe
2025
