SIMANFOR cloud Decision Suppo Sys em: S uc u e, con en ,
and applica ions
F. B a o
*
, C. O d´
o˜
nez , A. V´
azquez-Veloso, S. Michalakopoulos
SMART Ecosys ems G oup, Depa amen o de P oducci´
on Vege al y Recu sos Fo es ales, Ins i u o Uni e si a io de In es igaci´
on en Ges i´
on Fo es al Sos enible (iuFOR),
ETS Ingenie ías Ag a ias, Uni e sidad de Valladolid, A da. de Mad id 57, 34004 Palencia, Spain
ARTICLE INFO
Keywo ds:
Fo es managemen simula o
G ow h and yield
Fo es modelling
Cloud se ice
Sil icul u e
Ecosys em se ices
SUMMARY
Technological p og ess in he las decades has d i en g ea ad ances in many ields o knowledge. A wide ange
o ools and se ices a e now a ailable and cons an ly e ol ing o handle as amoun s o a ailable da a as well
as he inc eased complexi y o eal-wo ld case s udies and analy ical al e na i es. Mos sec o s ha e emb aced
new me hodologies o p o ide solu ions o hei p oblems, and he o es y sec o is no excep ion. Impo an
s eps ha e been aken o upda e he o es y sec o and in oduce new la ge-scale expe imen al designs, digi al
ools and mo e ex ensi e o es y da abases. Howe e , assimila ion o his p og ess by o es manage s emains
la gely pending. The mo e specialized echnical knowledge and compu ing skills equi ed o use his new gen-
e a ion o ools cons i u es a known ba ie o up ake. In his wo k, we p esen he SIMANFOR cloud-based
Decision Suppo Sys em se ice o simula ing o es managemen al e na i es. I s e olu ion, in e nal s uc-
u e and po en ial applica ions a e desc ibed. A case s udy was de eloped o demons a e simula o pe o mance
unde di e se managemen scena ios and highligh he bene i s o his ool o o es manage s. SIMANFOR cloud
se ices a e ee and can be accessed a www.siman o .es.
1. In oduc ion
Modelling is a c ucial complemen o obse a ional and expe i-
men al da a in o es science. The e m ‘model’ is o en used o desc ibe
h ee sepa a e concep s: pla o ms, models and pa ame e iza ions. To
cla i y, pla o ms a e so wa e ools o implemen ing models and hei
pa ame e iza ions. They a e no associa ed wi h a pa icula model bu
can e lec a speci ic modelling app oach: empi ical (P e zsch e al.,
2002), p ocess-based (G acia e al., 2003, 1999), hyb id (Ka ki e al.,
2023; Landsbe g and Wa ing, 1997), e c. Models a e abs ac ions o
o es dynamics a di e en le els ( ee, size-class, s and). These a e
ep esen ed by a se o equa ions, ules, and decisions ha enable use s
o o ecas o es dynamics. Local pa ame e iza ions a e adap a ions o
models o speci ic loca ions and species composi ions. The pla o ms
gene a e ep esen a i e ables, ee lis s, and g aphs o use by manage s
in ope a ional o es y. These ools ha e been in use o some ime and
a e e ol ing alongside o es managemen needs, o suppo
decision-making p ocesses.
Fo es managemen in ol es a g ea a ie y and quan i y o o es
in en o y da a (expe imen al plo s, LiDAR poin clouds, sa elli e da a,
e c.) and simula ion da a (g ow h, yield, clima e change scena ios, e c.).
The da a can be aw o cu a ed, de i ed om obse a ion/expe imen-
a ion o gene a ed h ough simula ions. I can be eco ded in di e en
ways (p i a e o public, wi h o wi hou emba go, linked open da a, e c.)
and s o ed locally o emo ely in di e se p op ie a y o open o ma s.
Da a managemen can in ol e people in di e en loca ions; i may
equi e wo kplace and emo e access as well as speci ic knowledge o
unde s and wha is behind he da a and a i e a he igh conclusions.
Typical aspec s include da a ga he ing, da a s o age, da a cu a ion,
p og amming, and ou pu gene a ion and in e p e a ion. He e, cloud
compu ing can acili a e he s o age and handling o massi e amoun s o
da a. Cloud compu ing has e ol ed o he poin ha we can ind many
examples o i in e e yday li e, and he o es y sec o ough o ha ness
i s po en ial.
A wide a ie y o da a is no mally used o s udy o es dynamics.
Pe manen plo s (PPs) allow esea che s o s udy he e olu ion o o es s
o e long pe iods. Some o he oldes plo s ha a e s ill in use da e back
o he 19 h cen u y (P e zsch, 2009). Expe imen al ne wo ks allow sci-
en is s o assess di e en e ec s on ees and s ands and obse e com-
mon beha io s among di e en loca ion g adien s (P e zsch e al., 2019;
Ve heyen e al., 2016). Fo sil icul u e ope a ions, hinning ials (Aldea
e al., 2017) a e a good example o how expe imen al ne wo ks seek o
* Co esponding au ho .
E-mail add ess: [email p o ec ed] (F. B a o).
Con en s lis s a ailable a ScienceDi ec
Ecological Modelling
jou nal homepage: www.else ie .com/loca e/ecolmodel
h ps://doi.o g/10.1016/j.ecolmodel.2024.110912
Recei ed 24 July 2024; Recei ed in e ised o m 10 Oc obe 2024; Accep ed 12 Oc obe 2024
Ecological Modelling 499 (2025) 110912
A ailable online 12 No embe 2024
0304-3800/© 2024 The Au ho s. Published by Else ie B.V. This is an open access a icle unde he CC BY-NC-ND license (
h p://c ea i ecommons.o g/licenses/by-
nc-nd/4.0/ ).
unde s and he e ec s o di e en ha es ypes. P o enance ials (Alía
e al., 2009) e alua e how ees om a ious o igins adap o speci ic
local condi ions. Some coun ies ha e also de eloped Na ional Fo es
In en o ies (Tomppo e al., 2010), which p o ide la ge amoun s o da a
o a a ie y o pu poses.
Inc easing conce n abou he e ec s o clima e change on o es
p oduc i i y and species dis ibu ion (Fe n´
andez-de-Una e al., 2015)
has shi ed he a en ion o he o es y sec o o mixed o es s, due o
hei highe esis ance and esilience (Toïgo e al., 2015). A Eu opean
iple ne wo k was implemen ed o s udy and compa e he single and
combined e ec s o clima e, and sil icul u e on pu e and mixed s ands o
se e al species (Del Río e al., 2017; Poeydeba e al., 2020). Such
expe imen al da a is being used o add ess ques ions ega ding p o-
duc i i y, compe i ion, hinning e ec s, species p opo ion ends, e c. I
allows o compa a i e analysis o pu e and mixed s ands, hinning in-
ensi ies and si e quali y. Howe e , o ans e his knowledge o o es
owne s and manage s, ools mus be implemen ed o simpli y usabili y.
Acco ding o Schieman and Fio do (1990), o es e s a e la e adop e s
o in o ma ion echnology ools. Se e al ini ia i es ha e been launched
o b idge he gap in ecen decades. Among hem a e he Sys em o
Ea h Obse a ions, Da a Access, P ocessing & Analysis o Land
Moni o ing (SEPAL) (FAO, 2022) and he Global Biodi e si y In o ma-
ion Facili y (GBIF) (GBIF.o g, 2023), global-scale examples o da abase
and da a u iliza ion esou ces. Da a managemen ools such as basi oR
(B a o e al., 2024; 2022) and Fo es Explo e (Vega-Go gojo e al.,
2022) o e solu ions a a egional scale. Se e al simula o op ions a e
also a ailable a egional and na ional le els. Simula o s a e digi al ools
ha can un di e en sil icul u al scena ios using o es -based inpu
da a and p o ide use ul in o ma ion o decision-making. They es ima e
o es me ics h ough models buil om a combina ion o equa ions
ha a e execu ed in he p ope calcula ion o de o es ima e a ious
o es me ics. The ma hema ical s uc u e o he models can a y o
accommoda e di e en loca ions and/o species; o i can be main ained
while he pa ame e s a e changed o each case s udy. Indeed, simula-
o s can be adap ed o speci ic loca ions and case s udies (P e zsch e al.,
2015). They o en implemen models applied o mixed o es s (Blanco
e al., 2015) and es ima ion o na u al dis u bances (Seidl e al., 2011),
p o iding new u ili ies o use s. Howe e , o es manage s s ill
encoun e ba ie s when i comes o adop ing hese echnologies. Easie
use in e ace wi h he pla o m, low ini ial ins alla ion and inpu da a
equi emen s, and mul iple applica ion op ions a e keys o expanding i s
use in suppo ing decision-making.
The main objec i e o his wo k is o show he oppo uni ies ha
cloud-based simula ions o e o he o es y communi y. He e, we
p esen SIMANFOR, a cloud-based o es managemen simula ion se -
ice and Decision Suppo Sys em (DSS) ha simula es a a ie y o o es
managemen al e na i es. We desc ibe SIMANFOR’s e olu ion om i s
o iginal o ma (B a o e al., 2012) o i s cu en s uc u e, he possi-
bili y o local o cloud-based use and he in eg a ion o he IBERO model
(B a o, 2005) as one o he many examples included in he pla o m. The
IBERO case s udy is p esen ed o illus a e he simula o ’s pe o mance
capabili ies and highligh i s bene i s o o es manage s.
2. SIMANFOR simula o
Al hough he idea behind SIMANFOR was concei ed ea lie , he i s
simula o e sion was o iginally de eloped in 2009 using he C# p o-
g amming language in .NET. F om he ou se , SIMANFOR has been
a ailable online (B a o e al., 2012). The o iginal e sion had p eloaded
he IBERO model a chi ec u e (B a o, 2005), designed o un single- ee
g ow h models independen o dis ance. O e ime, new pa ame iza-
ions we e included unde he same model s uc u e and he SILVES
model a chi ec u e (Del Río e al., 2005; Del Río and Mon e o, 2011)
was also implemen ed o un s and g ow h models. Addi ional
ecosys em se ice modules linked o hose models and hei pa ame-
iza ions we e implemen ed, such as ee c own me ics (Liza alde
e al., 2004), ee biomass (Ruiz-Peinado e al., 2012, 2011), ca bon
con en (Mon e o, 2005), and mush oom p oduc i i y (De La Pa a Pe al
e al., 2017; He e o e al., 2019).
In 2020, SIMANFOR was ew i en in he Py hon p og amming
language (Van Rossum and D ake, 2009), which imp o ed pe o mance
and added new unc ionali ies. Equa ions dependen on ee dis ance
(Uzquiano e al., 2021), clima e-dependen g ow h models o ungi
p oduc i i y in Medi e anean sc ublands (He n´
andez-Rod íguez e al.,
2015) and clima e-dependen g ow h models o mixed s ands
(Rod íguez de P ado, 2022) we e implemen ed in he new e sion. Case
s udies o ecosys em se ices p oduc ion (Rod íguez de P ado e al.,
2023; V´
azquez-Veloso e al., 2024; 2022), in eg a ion o mixed-s and
models (B a o and V´
azquez-Veloso, 2024; Rod íguez de P ado e al.,
2023) and model e alua ion and alida ion (V´
azquez-Veloso e al.,
2023) we e also ca ied ou wi h he SIMANFOR simula o .
The so wa e a chi ec u e has a modula design o sepa a e web
in e ace om he simula o co e. Wi h his app oach, he simula o can
be used in he cloud, locally o ia a high-pe o mance compu e when
needed (Fig. 1).
In he SIMANFOR a chi ec u e, ou so wa e componen s un in
Docke con aine s (Me kel, 2014), a echnology widely used in mode n
so wa e de elopmen o acili a e he design and deploymen o
modula componen s o use in di e en ope a ing sys ems and cloud
en i onmen s. The Docke con aine s un on i ual se e s whe eas he
i h componen uns on a physical se e .
The i e componen s (Fig. 1), a e:
•A MongoDB (Banke e al., 2016) da abase Docke con aine ha
s o es all he applica ion en i ies, such as o es in en o ies, sil i-
cul u e scena ios, models and use s.
•The backend Node.js +Exp ess (Exp ess JS, 2024; Node JS, 2024)
con aine ha manages he applica ion’s business logic a he co e o
he applica ion. I communica es wi h he MongoDB da abase (using
Ja asc ip ), he on end ( ia Rep esen a ional S a e T ans e
Applica ion P og amming In e ace (REST API)), and he simula o
se e ( ia he Secu e Shell p o ocol (SSH)).
•The on end is an Angula 9 (So, 2018) Docke con aine ha e-
cei es use inpu , in e ac s wi h he backend ia REST API and dis-
plays esul s and in o ma ion o he use .
•The ou h componen is an in e se Nginx p oxy (Soni, 2016), which
ou es use eques s o he on end o backend acco ding o he
eques and manages he applica ion’s web in e ac ions.
•The simula o cu en ly uns on adi ional clien -se e a chi ec-
u e, ecei ing SSH commands om he backend and scheduling and
moni o ing jobs using SLURM (Yoo e al., 2003). The simula o is
cu en ly being modi ied and buil as a Docke se ice and will un in
i s own con aine . The abili y o execu e simula ions on he supe -
compu e a Cal´
endula HPC (SCAYLE, 2019) in he o m o SLURM
jobs will emain as an op ion o jobs wi h hund eds o housands o
plo s and ees.
Despi e he p e ious implemen a ion changes, i s concep ual s uc-
u e as a simula o has emained he same. SIMANFOR is di ided in o
h ee main modules ha ha e emained cons an since he i s e sion:
ini ializa ion, p ojec ion, and hinning (Fig. 2). To s a he simula ion, a
o es in en o y and a sil icul u e scena io summa izing he ime p o-
jec ion and hinning ac i i ies mus be p o ided as inpu da a. Once a
simula ion s a s, he inpu da a is ead by he sys em and he ini iali-
za ion p ocess begins, in which missing alues o he ini ial in en o y
a e impu ed. A e ha , he simula o pe o ms each o he s eps sum-
ma ized in he sil icul u e scena io ollowing he o iginal o de . When a
ime p ojec ion is selec ed, hen equa ions conce ning mo ali y,
g ow h, and ing ow h a e execu ed and ee and s and me ics a e
upda ed. When a hinning e en is selec ed, hinning is applied ac-
co ding o he use hinning equi emen s o ype (sys ema ic, om
abo e/below), in ensi y, and c i e ia (pe cen age o ees ex ac ed
F. B a o e al.
Ecological Modelling 499 (2025) 110912
2
based on s and densi y, basal a ea, o olume). Finally, ee and s and
me ics a e upda ed. When all s eps in he scena io ha e been execu ed,
a unique ile is gene a ed o each plo . I includes a simula ion summa y
wi h de ailed ee and plo in o ma ion o each s ep o he scena io.
3. SIMANFOR s and-alone simula o
The abili y o un SIMANFOR locally is a de elopmen al b eak-
h ough. De elope s can now access e e y model s uc u e implemen ed
by he simula o and add new pa ame iza ions, implemen new model
s uc u es, o include addi ional calcula ions such as ecosys em se ices.
The local e sion also educes he e o o es ing and debugging new
implemen a ions, while ensu ing he s abili y o cloud se ice upda es.
Ex e nal use s in e es ed in con ibu ing wi h new models o pa-
ame iza ions can do so by p o iding: (1) a model name and desc ip-
ion; (2) he model s uc u e and equa ions o be implemen ed; (3)
model inpu equi emen s; (4) c edi s (au ho ship and ci a ion); and (5)
con ac de ails. The in o ma ion is summa ized in a model desc ip ion
shee a ailable o use s as a guide ha de elope s can ollow o p o-
g am a new model o pa ame iza ion, acco ding o he wo k low shown
in Fig. 3. In a nu shell, con ibu ing use s iden i y a model o pa ame-
e iza ion no cu en ly suppo ed by he simula o and submi a eques
o he de elopmen eam o i s inclusion. Wi h he eques , hey p o ide
documen a ion o he model and a case s udy wi h da a and expec ed
ou pu . Ou de elope s implemen i in he SIMANFOR s and-alone
simula o , hen in e es ed use s es he model and epo ing bugs o
he main de elope . Tha debugging p ocess con inues un il he com-
pu a ions a e execu ed smoo hly, p edic ion esul s a e sa is ac o y, and
a s able e sion o he model/pa ame iza ion is ob ained. This can hen
be included in he SIMANFOR cloud and linked o documen a ion and
es da a.
When he numbe o plo s included in he in en o y eaches a ce ain
le el and/o he numbe o scena ios o simula e is e y la ge, he
compu ing equi emen s inc ease and become un easible o un on a
pe sonal compu e . Howe e , la ge-scale simula ions can be launched in
a high-pe o mance compu ing (HPC) en i onmen using he SIMAN-
FOR local simula o . Use s can con ac he SIMANFOR echnical eam
and a ange o un simula ions on he iuFOR – Uni e si y o Valladolid
supe compu e . Simula ions wi h e en g ea e da a and/o p ocessing
equi emen s can also be un a Cal´
endula HPC, he High-pe o mance
Compu ing Cen e o Cas illa y Le´
on (SCAYLE, 2019). These cen e s
ha e been used in p io published wo ks de eloped wi h SIMANFOR
(Rod íguez de P ado e al., 2023; V´
azquez-Veloso e al., 2023b).
4. SIMANFOR cloud se ice
The SIMANFOR web in e ace is a ailable in mul iple languages
(English, Spanish, F ench, Po uguese, Vie namese, Galician, and Bas-
que). I s help sys em con ains se e al esou ces and use manuals in
English and Spanish. Use equi emen s include an in e ne connec ion,
a web b owse ( he pla o m has been es ed on Mic oso Edge, Google
Ch ome, Ch omium, Mozilla Fi e ox, and Sa a i b owse s) and a
sp eadshee p og am o da a managemen (Mic oso Excel, Lib eO ice
Calc o OpenO ice Calc).
Two di e en oles exis o he SIMANFOR cloud: adminis a o and
use . Adminis a o s a e esponsible o managing he sys em and
au ho izing p i ileges o o he use s. Adminis a o s upload models
al eady de eloped (as explained be o e), p o ide eliable documen a-
ion and es da a, and ensu e model accu acy and p ope pe o mance.
Use s can upload in en o y da a, check a ailable models and pa ame-
iza ions, selec he pa ame iza ion ha bes i s he in en o y da a
and simula e o es managemen scena ios. Use s a e esponsible o he
adequacy and accu acy o hei managemen scena ios, he in en o y
da a hey use and he models hey choose o hei simula ions.
The in en o y da a p o ided o he sys em is a c i ical poin , as i
mus comply wi h he simula o ’s equi emen s. A ailable models ha e
sample da ase s o se e as a guide, and in en o y empla es a e a ail-
able o help educe use e o . Fo ield da a collec ion, as an example,
he T eeCollec And oid (h ps://www.and oid.com/) app was de el-
oped o easily eco d ee in o ma ion and upload i o he SIMANFOR
cloud (B a o e al., 2017). Cu en e o s a e ocused on he in e ace
be ween Fo es Explo e , an in e ac i e ool o explo ing he Spanish
Na ional Fo es y In en o y (Vega-Go gojo e al., 2022), and he
SIMANFOR cloud. Wi h easy in ui i e s eps, use s will soon be able o
p epa e hei SIMANFOR- eady in en o y on Fo es Explo e by selec -
ing hei a eas o in e es (p o ince, plo , o polygon), applying il e
Fig. 1. SIMANFOR web in e ace and simula o a chi ec u e.
F. B a o e al.
Ecological Modelling 499 (2025) 110912
3
Fig. 2. SIMANFOR simula o s uc u e.
F. B a o e al.
Ecological Modelling 499 (2025) 110912
4
Fig. 3. Wo k low o model/pa ame iza ion de elopmen on SIMANFOR, ep esen ing a case o including a new model. The same wo k low applies o new
pa ame iza ions.
F. B a o e al.
Ecological Modelling 499 (2025) 110912
5
c i e ia (species, age…) and uploading he esul s o he SIMANFOR
cloud.
Scena io managemen is ano he key aspec o he cloud se ice.
Use s can selec hei p e iously uploaded in en o y da a, he model and
pa ame iza ion ha be e i s hei equi emen s, and easily de elop a
sil icul u e scena io by gene a ing ime p ojec ions and hinning ac ions
un il he desi ed o a ion pe iod. An addi ional se ice is a ailable ia
SMARTELO APP (V´
azquez-Veloso e al., 2023a), an And oid applica ion
ha allows use s o plan ha es s di ec ly in he ield while explo ing
ee me ics. Fo es manage s can gene a e a sil icul u e plan and
compa e i on he websi e wi h al e na i es gene a ed by hemsel es o
o he s. Bo h he in en o y and he de ails o he hinned ees a e sen o
he SIMANFOR websi e, so use s can ini ia e al e na i e u u e sil i-
cul u al scena ios ollowing ini ial selec i e hinning in he ield.
The simula ion esul s a e ob ained a e unning each scena io. An
Excel ile is gene a ed o each plo and scena io simula ed, as he in-
en o y da a can include se e al plo s. Each Excel ile includes a yield
able as a summa y o he scena io; a s and shee , whe e all he a iables
a e calcula ed o each scena io s ep; ee shee s (one shee pe s ep)
wi h upda ed in o ma ion o each ee; a desc ip ion shee , whe e in-
o ma ion abou he model is p o ided; and a me ada a shee , whe e all
he a iables shown in he ile a e explained. A new unc ionali y is
cu en ly being de eloped ha will allow use s o isualize hei esul s
as dynamic g aphs de eloped wi h he Shiny R lib a y (Chang e al.,
2015). This acili a es easy g aphic explo a ion o esul s and compa i-
son wi h esul s om o he sil icul u e ac ions.
Finally, documen a ion has been de eloped o each aspec o he
simula o , o guide and suppo SIMANFOR cloud use s. Manuals, es
da ase s, model desc ip ions, ou pu examples and explana o y ideos
a e a ailable a : h ps://gi hub.com/siman o (SIMANFOR, 2022).
5. SIMANFOR applica ions
SIMANFOR is being used in h ee main a eas: educa ion, esea ch,
and o es managemen .
5.1. Educa ion
SIMANFOR is an inc edibly e sa ile educa ional ool o o es y
s uden s, o es manage s and o es owne s. Wi h a ange o usabili y
op ions, i allows use s o explo e he e ec s o sil icul u e, clima e,
species mix u es, and o he scena ios. SIMANFOR simula ion ou pu
makes i possible o assess nume ous ecosys em se ices and es s a-
egies ela ed o wood and non-wood esou ce p oduc ion, allome ies,
species di e si y and mixed-s and p opo ions (Rod íguez de P ado
e al., 2023; V´
azquez-Veloso e al., 2024). O e all, SIMANFOR p o ides
an imme si e expe ience o help use s ind he bes sil icul u al s a egy
and bols e hei knowledge abou o es modelling. The new,
use - iendly cloud in e ace p o ides a be e use expe ience, while
associa ed apps such as T eeCollec (B a o e al., 2017) and SMARTELO
APP (V´
azquez-Veloso e al., 2023a) enhance he simula o ’s usabili y
and acili a e p ac ical aining oppo uni ies in he ield.
5.2. Resea ch
SIMANFOR was bo n in a scien i ic en i onmen and has been a
aluable esea ch ool in many a eas since i s c ea ion. I o e s p ac ical
suppo o unde s anding ee- and s and-le el o es dynamics,
enabling di ec compa isons o p oduc i i y, clima e, and sil icul u e
scena ios. SIMANFOR is especially help ul o s udying mixed o es s
and mix u es ha ha e no ye been included in expe imen al ials. I s
simula ions can also p o ide ini ial insigh s abou s and dynamics when
expe imen al da a is lacking (B a o and V´
azquez-Veloso, 2024).
SIMANFOR can gene a e comp ehensi e g ow h and yield da a o wood
and non-wood esou ces, such as ungi (De La Pa a Pe al e al., 2017)
and pine nu p oduc ion (V´
azquez-Veloso e al., 2022) in di e en
scena ios, hus u he ing s udy o he dynamics in ol ed. O he
esea ch e o s ha e explo ed ca bon con en in ela ion o di e se
sil icul u al p ac ices (Ma ín A iza e al., 2017) o in mixed s ands
unde a ying clima e scena ios (Rod íguez de P ado e al., 2023).
Pionee ing s eps ha e al eady been aken o de elop me hodologies o
model e alua ion and alida ion (V´
azquez-Veloso e al., 2023). Fu u e
esea ch lines will in eg a e LiDAR me ics and models o expand use
and esea ch capabili ies.
5.3. Fo es managemen
As a simula o , SIMANFOR excels in i s applicabili y o o es man-
agemen , equipping and empowe ing use s wi h a wide a ie y o
sil icul u al al e na i es o assessmen and compa ison. Whe he a
ee, size-class, o s and le el, SIMANFOR can simula e g ow h and yield
o di e se empo al scopes. I acili a es s a ic yield simula ions along
wi h ex ensi e managemen plans spanning o e a cen u y, and suppo s
decision-making by enabling use s o assess a ange o al e na i es.
Unlike ea lie models ha we e p ima ily designed o pu e s ands,
SIMANFOR includes mixed models ha open new possibili ies o
explo ing managemen s a egies in scena ios o limi ed managemen
expe ience, heigh ened complexi y, and al e ed o es dynamics
(P e zsch and Schü z, 2014). To illus a e hese capabili ies, a case s udy
is p esen ed in he ollowing sec ions. I desc ibes how he IBERO model
was in eg a ed in o SIMANFOR, and s udies s and e olu ion unde
di e en managemen scena ios.
6. IBERO model in eg a ion
IBERO is an indi idual- ee g ow h model independen o dis ance
de eloped by B a o (2005). I was o iginally pa ame ized o Pinus
pinas e mesogeensis and Pinus syl es is e en-aged pu e s ands. IBERO
acili a es i e-yea p ojec ions (longe p ojec ions can be made wi h
successi e i e-yea p ojec ions) and inco po a es di e en modules ha
a e easily in eg a ed in o a SIMANFOR wo k low (Fig. 4; see also B a o
(2005) o de ails).
Once a simula ion s a s and da a is loaded, he ini ializa ion p ocess
begins. He e, he pa ame iza ion ha bes i s each plo included in he
in en o y is selec ed ( o example, Pinus pinas e and Pinus syl es is ha e
di e en pa ame iza ions). P oduc i i y, d i en by Si e Index da a
(B a o and Mon e o, 2001; B a o-O iedo e al., 2004), is calcula ed and
does no a y du ing he en i e simula ion o each plo . T ee me ics a e
hen calcula ed and de aul me ics such as basal a ea o bal (Wyko
e al., 1982) a e impu ed whe e in o ma ion is missing. Species-speci ic
me ics such as ee heigh (H/D), c own and olume (Liza alde, 2008),
me chan able wood olumes (Rod íguez, 2009), biomass (Ruiz-Peinado
e al., 2012, 2011), ca bon con en (Cas a˜
no-San ama ía and B a o,
2012; Mon e o, 2005), and mush oom p oduc i i y (He e o e al.,
2019; S´
anchez-Gonz´
alez e al., 2019) a e also calcula ed. A e ha ,
common plo me ics like s and densi y and dominan heigh , along wi h
me ics de i ed om he p e ious ee calcula ions such as s and olume
and biomass, a e impu ed in he plo in en o y.
Once ini ializa ion has inished, he simula o i e a es he s eps in he
scena io and execu es each one in he o de p o ided. I he nex s ep is a
p ojec ion, hen he su i al (B a o-O iedo e al., 2006), g ow h
(Liza alde, 2008), and ing ow h (B a o e al., 2008) modules a e ac i-
a ed o calcula e ee su i al p obabili y (Ps), diame e and heigh
inc emen (DBHi and Hi), ing ow h p obabili y (Pi), and basal a ea, all
o which a e hen inco po a ed o he espec i e s and (BAi). T ee and
plo me ics a e hen upda ed using he same equa ions. I he nex s ep
o he simula ion is hinning, hen he hinning module is ac i a ed ac-
co ding o use ins uc ions. T ee and s and me ics a e hen upda ed as
be o e.
Once he scena io has inished, he ou pu is w i en o a ile, along
wi h he in o ma ion pe aining o he model and he pa ame iza ion
used o pe o m he simula ion.
F. B a o e al.
Ecological Modelling 499 (2025) 110912
6
7. Case s udy
To illus a e how he SIMANFOR simula o pe o ms, a sil icul u al
simula ion was un using in en o y da a, he IBERO-PT model (IBERO
model pa ame iza ion o Pinus pinas e ), and he SIMANFOR cloud
se ice.
Fig. 4. In eg a ion o he IBERO model s uc u e in o he SIMANFOR simula o . G ey on e e s o speci ic componen s o he IBERO model ha a e explained in
he ex .
F. B a o e al.
Ecological Modelling 499 (2025) 110912
7
7.1. Da a
Ini ial da a comes om i e pe manen Pinus pinas e plo s loca ed in
he Sou he n Ibe ian Range (Spain). The plo s we e es ablished in 2003
(see Table 1) and emeasu ed in 2008 and 2013.
7.2. Sil icul u e scena ios
Fou sil icul u al scena ios we e de ined by combining hinning ype
( om below o om abo e), and hinning o a ion (e e y 10 o e e y 15
yea s). A con ol (no hinning) simula ion was also pe o med o
benchma k he esul s. In all cases, hinning in ensi y was se o 25%
educ ion in basal a ea. The p ojec ion ime span o he s and was om
he ini ial plo age o 81–83 yea s old, depending on he ini ial s and age.
A summa y is shown in Table 2, and mo e de ailed in o ma ion abou
each sil icul u al scena io is a ailable in Appendix A.
7.3. Resul s
The esul s o he main s and me ics a e shown in Fig. 5. Bo h
densi y (N) and basal a ea (G) show highe alues when no sil icul u e is
applied o he s and, while dominan heigh (Ho) as he quad a ic mean
diame e (dg) o scena ios 1 and 2 in which hinning om abo e is
applied, is lowe a he end o he simula ion. Signi ican di e ences
based on hinning c i e ia a e also no iceable. While scena ios 1 and 2
( hinning om below) show highe dg alues om he ou se due o he
emo al o smalle ees, Ho was consis en ly highe du ing he simu-
la ion un il 10 yea s ago, when i was su passed by Scena io 3 ( hinning
om abo e e e y 10 yea s). S and densi y was highe du ing he en i e
simula ion in scena ios 3 and 4 ( hinning om abo e) because he
hinning objec i e o 25% o G is easily eached when bigge ees a e
emo ed. G is also consis en ly highe in hose scena ios, hough no
compa ed o he e e ence Scena io 0.
The ime be ween hinnings also signi ican ly a ec ed he ou comes.
N p esen s lowe alues when hinning om below is ca ied ou e e y
10 yea s (Scena io 1) compa ed o 15 yea s (Scena io 2), as a mo e
in ense ex ac ion egime is being applied. Howe e , he opposi e
beha io appea s when compa ing hinning om abo e. Highe N
alues a e obse ed in Scena io 3, whe e hinning is pe o med e e y 10
yea s. Ho clea ly p esen s highe alues when hinning is pe o med
e e y 10 yea s wi h bo h hinning ypes, which is likely ela ed o he
highe compe i ion be ween emaining indi iduals. Di e ences in dg
alues a e also signi ican when hinning is applied om below (sce-
na ios 1 and 2) bu show no disce nable di e ences when applied om
abo e (scena ios 3 and 4). Di e ences in G a e no ema kable among
any o he hinning scena ios.
To accompany hese esul s, yield ables a e a ailable in Appendix B
and complemen a y g aphs o olume, biomass, and ca bon con en a e
a ailable in Appendix C. The o iginal esul s con ained ex ensi e ee-
and s and-le el me ics da a, which can be ound in he complemen a y
da a a ached o his a icle.
8. Discussion
The IBERO case s udy illus a es SIMANFOR’s po en ial o simu-
la ing and compa ing o es managemen al e na i es. Al hough only
common s and a iables we e conside ed and compa ed, SIMANFOR
gene a ed a e y la ge numbe o me ics a bo h ee and s and le el.
This demons a es i s capabili ies o esponding o a a ie y o eques s,
including mo e complex ones. Addi ionally, he sepa a ion be ween he
simula o and he cloud se ice p o ides g ea e lexibili y o he
de elopmen o new unc ionali ies and he inclusion o mo e me ics.
The SIMANFOR echnical eam handles new implemen a ions ha use s
may eques .
While IBERO was he i s model s uc u e in eg a ed in o SIMAN-
FOR, new models wi h di e en s uc u es ha e since been imple-
men ed. In cases whe e only s and in o ma ion is a ailable and ee da a
is missing, s and models can be applied on he pla o m. When he aim is
o manage mixed o es s o simula e he e ec s o a ious clima e sce-
na ios, speci ic models a e a ailable on he pla o m o suppo hese
si ua ions (implemen a ion is desc ibed in B a o and V´
azquez-Veloso
(2024)). The mixed- o es model s uc u e includes pa ame iza ions o
he 29 mos ex ensi ely occu ing wo-species combina ions in Spain
and makes i possible o use s o compa e sil icul u e scena ios and
hei e ec s unde u u e clima e condi ions. A case s udy compa ing
ca bon seques a ion in ou Pinus syl es is mixed s ands is a ailable in
Rod íguez de P ado e al. (2023). Thus, he possibili y o implemen ing
new models, as he CAPSIS simula o does (Du ou -Kowalski e al.,
2012), gi es SIMANFOR an ad an age o e FVS (C ooks on and Dixon,
2005) and simila al e na i es. SIMANFOR’s capaci y o in eg a e pa-
ame e iza ions o di e en species and loca ions, o e en di e en
coun ies like Spain and Mexico, u he enhance he usabili y o his
simula o .
As a eely accessible ool, SIMANFOR enables o es manage s and
o es scien is s o implemen , es and un g ow h and yield models.
Howe e , h oughou he SIMANFOR de elopmen p ocess, ce ain
ba ie s ha e been appa en : Fo a s a , ope a ional o es e s a e un-
amilia wi h cloud compu ing and he e o e eluc an o apply his
echnology. SIMANFOR is e y lexible, which some use s may pe cei e
as o e ly complica ed. Also, he DSS elies on he con e gence o good
models, good da a and good simula ions. Finally, he ime and e o
equi ed o unde s and how SIMANFOR wo ks may be highe han
ini ially expec ed. De elope s and use s should keep hese ba ie s in
mind when designing o adop ing cloud-based DSS o acili a ing he use
o his echnology in ope a ional o es y. To o e come hese obs acles,
we a e wo king o: keep he SIMANFOR cloud expe ience as simple and
use -o ien ed as possible; de elop ies wi h he o es y communi y (we
a e de eloping ope a ional yield ables o engage hem); keep he cloud
in e ace clea and simple; p o ide adequa e help and suppo o local
p oblems along wi h accu a e in o ma ion o public discussion. Though
ime cons ain s and economic ba ie s end o hinde such ac ions, we
belie e hem o be he mos in luen ial ac i i ies o encou aging he
adop ion o cu ing-edge echnology. Fo example, he use in e ace
educes he e o equi ed o access and implemen SIMANFOR (see also
C ooks on and Dison, 2005; Du o -Kowalski e al., 2012). Simplici y and
cloud access ha elimina es ins alla ion equi emen s gi e SIMANFOR
an edge o e al e na i es wi hou use in e ace, such as he SIMO
Table 1
Summa y o he in en o y da a. N is he s and densi y; dg is he quad a ic mean
diame e ; G is he s and basal a ea; Ho is he s and dominan heigh ; and SI is he
Si e Index a 80 yea s old.
Plo ID Age N dg G Ho SI
(yea s) ( ees/ha) (cm) (m
2
/ha) (m) (m)
2,642,306 26 1510 15.1 27.2 10.8 22.1
2,644,107 29 2451 10.7 21.9 6.5 15.3
2,642,105 32 509 26.9 28.9 13.2 22.6
2,642,108 32 912 20.2 29.3 10.9 19.9
2,619,104 33 1252 13.7 18.4 8.0 16.1
Table 2
Summa y o he scena ios simula ed o speci ic hinning ypes and o a ions.
Thinning in ensi y emained ixed a 25% in basal a ea o all in e en ions.
Scena io code Thinning ype Thinning o a ion
0 None None
1 om below 10 yea s
2 om below 15 yea s
3 om abo e 10 yea s
4 om abo e 15 yea s
F. B a o e al.
Ecological Modelling 499 (2025) 110912
8
simula o (Rasinm¨
aki e al., 2009), R lib a ies o SiT ee
(An ´
on-Fe n´
andez and As up, 2022), which equi e p og amming skills
o allome ic knowledge (F ank e al., 2023).
Addi ional se ices ha a e al eady in eg a ed in o he simula o o
o e come usabili y ba ie s include low compu a ional equi emen s,
on-demand access o HPC se ices, acili ies o de elop and manage
o es in en o ies, he possibili y o simula ing mul i-plo da a se s
simul aneously and ailo ing he ou pu o speci ic equi emen s, and
he abili y o include and es new models, pa ame iza ions o me ics
as needed o he end use s. The ‘in e ne o hings’ (IoT) can also play a
c ucial ole in collec ing o es da a (Bo e al., 2011) o p ocess-based
modelling. De ailed da a can be in eg a ed in o spa ially explici
g ow h models o imp o e p edic ions a a ious ime and spa ial le els.
Fu he wo k is needed o implemen IoT ou comes as pa o he
SIMANFOR pla o m.
Fo es moni o ing and o ecas ing h ough cloud compu ing pla -
o ms enhances he pe o mance o o es scien is s and manage s
h ough ools such as Google Ea h Engine (GEE) o SIMANFOR by
accele a ing da a collec ion, da a p ocessing, knowledge gene a ion and
ope a ional applica ions. To u he in eg a e cloud-based compu ing
in o o es y, Han e al. (2023) has p oposed a dis ibu ed s o age model
o help manage s e alua e o es p ocesses ha impac ca bon
seques a ion.
Finally, while many esou ces ha e been de eloped, new p ojec s a e
unde way o p o ide a be e expe ience and mo e ools o SIMANFOR
use s. We o esee ha use s will soon be able o iew simula ion esul s
in a g aphically pleasing and illus a i e manne using Shiny R so wa e
(Chang e al., 2015). Simila ly, we a e wo king o de elop mo e in ui i e
c ea ion and edi ing o sil icul u e scena ios, while also packaging he
s and-alone simula o in o a Docke con aine mic ose ice. F om he
use suppo side, he manuals a e cons an ly being upda ed and in o-
duc o y ideos will educe he ime spen lea ning o use he simula o .
F om he echnical side, u u e upda es will include new me ics ha
p o ide mo e de ails abou use sil icul u e, along wi h new models and
pa ame iza ions o di e en species and loca ions. We a e also wo k-
ing owa ds he implemen a ion o LiDAR me ics and models o expand
usabili y. Mo e ambi ious wo ks in he pipeline include he in eg a ion
o Fo es Explo e as desc ibed in Gim´
enez-Ga cía e al. (2024), which
will allow use s o isually selec hei plo s and c ea e in en o ies ha
can be uploaded di ec ly o he SIMANFOR cloud. We also hope o
in oduce a low-code applica ion add-on o easie inclusion o a new
model o pa ame e iza ion o an exis ing one, along wi h cus omizable
in e ace ea u es ha allow use s o pe sonalize he look and eel o he
applica ion.
9. Conclusion
This pape p esen s he SIMANFOR o es managemen simula o , i s
e olu ion and he s uc u e o he simula o and cloud se ice, which a e
eely a ailable a www.siman o .es. SIMANFOR cloud is a suppo ool
o o es managemen ha p o ides a la ge amoun o in o ma ion wi h
li le use e o h ough a simple, use - iendly in e ace. I s applica-
bili y has been amply demons a ed in educa ion, esea ch and o es
managemen . SIMANFOR is unde con inuous de elopmen and new
ea u es a e cons an ly being inco po a ed. Fu he mo e, SIMANFOR
can in eg a e models and pa ame e iza ions om di e en ecosys ems
a ound he wo ld. I s simula ions a e suppo ed by a se e ha can be
scaled up o accommoda e u u e demands.
Fig. 5. G aph compila ion wi h esul s o he main s and me ics. The e olu ion o s and densi y ( op-le ), dominan heigh ( op- igh ), quad a ic mean diame e
(bo om-le ) and basal a ea (bo om- igh ) a e shown wi h he a e aged esul s o he 5 plo s s udied. Each scena io is ep esen ed by one colo , as shown in he
legend. The peak ha appea s a unde 30 yea s in all he g aphs esul s om he di e en s and ini ial ages, as s and esul s a e a e aged and no ini ial in o ma ion is
a ailable in some cases.
F. B a o e al.
Ecological Modelling 499 (2025) 110912
9
