Anne e Achieng’
EFFECT OF LAND USE LAND COVER
CHANGES ON CARBON
SEQUESTRATION IN GERMANY
ii
EFFECT OF LAND USE LAND COVER CHANGES
ON CARBON SEQUESTRATION IN GERMANY
Disse a ion supe ised by:
Ped o da Cos a B i o Cab al, PhD
Co-supe ised by:
Judi h Ve s egen, PhD
Se gi T illes Oli e , PhD
Feb ua y 2020
iii
ACKNOWLEDGEMENTS
My i s g a i ude goes o my pa en s o enabling me o be he e o pu sue his mas e ’s
p og am. Secondly, I would like o hank my supe iso s o hei indispensable ad ice
and ema ks o wi hou hem I would no ha e been able o comple e his hesis.
Las ly, hanks a e due o he depa men o Geospa ial Technologies a NOVA IMS,
including all my p o esso s. Fo wi hou hei assis ance, suppo and he knowledge
ha hey ha e impa ed upon me h oughou his mas e ’s p og am, I would no ha e
been able o comple e his hesis.
i
DECLARATION OF ORIGINALITY
I decla e ha he wo k desc ibed in his documen is my own and no om someone
else. All he assis ance I ha e ecei ed om o he people is duly acknowledged and
all he sou ces (published o no published) a e e e enced.
This wo k has no been p e iously e alua ed o submi ed o NOVA In o ma ion
Managemen School o elsewhe e.
[Lisboa], [17/02/2020]
[Anne e Achieng’]
EFFECT OF LAND USE LAND COVER CHANGES
ON CARBON SEQUESTRATION IN GERMANY
ABSTRACT
Using ca bon seques a ion as an indica o o en i onmen al heal h, i is possible o
assess whe he a coun y is on i s way o achie ing ca bon neu ali y in he Land Use,
Land-Use Change and Fo es y (LULUCF) sec o . A g ea deal o esea ch has been
conduc ed o ind ou whe he he e is a ela ionship be ween LULC and ca bon
seques a ion. In his pape we explo e se e al scena ios and compa e how much
ca bon would be s o ed unde each o hem. In addi ion, his esea ch aims o ind ou
how bes he LULUCF sec o can con ibu e owa ds a coun y’s goals in achie ing
ca bon neu ali y. This was conduc ed using wo models; In eg a ed Valua ion o
Ecosys em Se ices and T ade-o s (InVEST) model, which calcula es he amoun o
ca bon s o ed in a landscape and Te Se ’s Land Change Modelle , which uses a
combina ion o neu al ne wo ks and CA Ma ko o p ojec u u e Land Use Land
Co e (LULC) scena ios. F om he documen a ion o he ca bon end o e he 28-
yea pe iod using he InVEST model, he s udy inds ha be ween he yea s o 1990
and 2018, he amoun o ca bon s o ed inc eased by 0.15%. Unde he Business as
Usual scena io p ojec ion he e is an inc ease o 0.22% by he yea 2048. In he
de elopmen scena io we see a dec ease o 0.96% and inally in he wo conse a ion
scena ios he ca bon s ock inc eases by 4.16% and 0.41% espec i ely. These esul s
sugges ha he scena io which would be mos bene icial o Ge many would be he
i s conse a ion scena io. The esul s o his s udy highligh he impo ance o he
LULUCF sec o in mi iga ing clima e change. The e o e, hey can be used o p o ide
in o med decision making o spa ial planne s and land managemen s akeholde s
du ing he de elopmen o u u e land use planning policies.

i
KEY WORDS:
Ca bon seques a ion
InVEST model
Geog aphic In o ma ion Sys ems
Te Se ’s Land Change Modelle
ii
ACRONYMS
BAU Business as Usual
CAP Clima e Ac ion Plan
CLC Co ine Land Co e
DEM Digi al Ele a ion Model
ES Ecosys em Se ices
FAO Food Aid O ganiza ion
GDP G oss Domes ic P oduc
GHG G eenhouse Gases
InVEST In eg a ed Valua ion o En i onmen al Se ices and T ade-o s
IPCC In e na ional Panel o Clima e Change
K o C Kilo ons o Ca bon
LCM Land Change Modelle
LULC Land Use Land Change
LULUCF Land Use, Land-Use Change and Fo es y
MLP Mul i-laye Pe cep on
Mg o C Megag ams o Ca bon
PPP Pu chasing Powe Pa y
REDD+ Reducing o Emissions om Fo es Deg ada ion and De o es a ion
UK Uni ed Kingdom
UNFCC Uni ed Na ions F amewo k Con en ion on Clima e Change
USGS Uni ed S a es Geological Su ey
WEC Wo ld Economic Fo um
iii
INDEX OF THE TEXT
ACKNOWLEDGEMENTS .....................................................................................................iii
DECLARATION OF ORIGINALITY .................................................................................... i
ABSTRACT .............................................................................................................................
KEY WORDS: ......................................................................................................................... i
ACRONYMS .......................................................................................................................... ii
INDEX OF THE TEXT ......................................................................................................... iii
INDEX OF TABLES ............................................................................................................... ix
INDEX OF FIGURES ..............................................................................................................x
1. In oduc ion ...................................................................................................................... 1
2. S udy a ea ......................................................................................................................... 5
3. Da a and me hods ............................................................................................................. 7
3.1 De ailing he p og ession o LULC changes in Ge many om 1990-2018 ................... 8
3.2 Ca bon s ock end calcula ion ....................................................................................... 8
3.3 P ojec ing u u e scena ios o ca bon seques a ion. ...................................................... 9
4. Resul s ............................................................................................................................ 14
4.1 LULC changes in Ge many om 1990-2018 .............................................................. 14
4.2 Ca bon seques a ion in Ge many om 1990-2018 ..................................................... 16
4.3 Modelling o u u e scena ios ...................................................................................... 18
5. Discussion ...................................................................................................................... 21
5.1 LULC changes and ca bon s ock in Ge many om 1990-2018 .................................. 21
5.2 Modelling o u u e scena ios ...................................................................................... 22
6. Conclusion ..................................................................................................................... 28
Bibliog aphic Re e ences ....................................................................................................... 29
ANNEXES ............................................................................................................................... 1
1. Reclassi ied LULC maps ............................................................................................. 1
2. Independen a iables .................................................................................................. 2
3. Ca bon s ock maps om modelled scena ios ............................................................... 3
ix
INDEX OF TABLES
Table 1: Ca bon pool alues o abo e g ound (Cab al e al., 2016) ....................................... 8
Table 2:Di ision o he land a ea in Hec a es among he ansi ion po en ials .................... 18
5
2. S udy a ea
Ge many, o mally known as he Fede al Republic o Ge many has a o al a ea o 357,
386 squa e kilome es (35.7386 million ha) and is loca ed in Cen al Wes e n Eu ope
a 51°N and 9°E. The coun y bo de s Denma k o he no h, Luxembou g, Belgium
and he Ne he lands o he wes , Poland and Czech Republic o he eas , Aus ia and
Swi ze land o he sou h, and las ly, F ance o he sou hwes . i is go e ned by a ede al
pa liamen a y epublic
Ge many has a popula ion o 83,686,264 making i he mos populous s a e in he
Eu opean Union. Nea ly hal o he popula ion is concen a ed be ween he wo king-
class popula ion (25-54 yea s) and he senio popula ion (65 yea s+) a 38.58% and
22.99% espec i ely (“Ge many Popula ion (2020) - Wo ldome e ,” n.d.).
I has a coas line o 2,389 kilome es and a mean ele a ion o 263 me es. Ge many’s
majo i e s a e he Rhine, Wese , Ode , and Elbe which low no hwa d, and he
Danube which has i s sou ce in he Black o es and lows eas wa d. Mos o he
coun y expe iences a empe a e seasonal clima e ha is domina ed by humid wes e ly
winds.
Economically, Ge many is he wo ld’s hi d la ges expo e o goods; hei main
expo s a e ehicles, machine y, chemical goods, elec onic p oduc s, elec ical
equipmen , pha maceu icals, anspo equipmen , basic me als, ood p oduc s, and
ubbe and plas ics. Ge many is he wo ld’s i h la ges economy by PPP (pu chasing
powe pa i y) and ou h by nominal GDP (g oss domes ic p oduc ) (“Eu ope ::
Ge many — The Wo ld Fac book - Cen al In elligence Agency,” n.d.).

6
Figu e 1: S udy a ea
7
3. Da a and me hods
The main da ase used o his esea ch consis ed o i e CLC maps sou ced om he
Cope nicus Land Moni o ing Se ice websi e (“CORINE Land Co e — Cope nicus
Land Moni o ing Se ice,” n.d.). These maps co e i e di e en ime pe iods o :
1990, 2000, 2006, 2012 and 2018. They ha e a esolu ion o 100 me es, a scale o
1:100,000 and a hema ic accu acy o abo e 85% wi h a minimum mapping uni o
25ha/100m.
The maps we e i s eclassi ied om he o iginal 44 classes o he 5 basic classes o
Land Use Land Co e (LULC) maps which a e: a i icial su aces, ag icul u al a eas,
o es and semi na u al a eas, we lands and wa e bodies. This is done because he
ca bon pool able, which is he second inpu o he InVEST (In eg a ed Valua ion o
Ecosys em Se ices and T ade-o s) model, con ains he 5 basic classes. In o de o
he model o un and p oduce alid esul s, he classes mus ma ch.
The ca bon pool able is a able con aining alues o ca bon s o ed in each o he
essen ial pools o each LULC class. The ou ca bon pools a e de ined by:
abo eg ound biomass which consis s o all li ing biomass occu ing abo e he soil,
belowg ound biomass which is all he li ing biomass in li e oo s, dead wood which
is all non-li ing woody biomass ha is no con ained in li e , soil o ganic ca bon
which is o ganic ca bon ound in mine al and o ganic soils (Un ccc, 2015). The sizes
o hese pools a e esponsible o ca bon s o age on land (Un ccc, 2015). Fo he
pu poses o his s udy, only he abo eg ound biomass was conside ed because he
s udy aims o analyse ca bon s o ed in he landscape.
The able used mus con ain he ollowing i e columns: lucode (unique iden i ie o
each LULC class in he CLC maps), LULC_name ( he names o he i e classes),
C_abo e ( he ca bon densi y con ained in abo eg ound biomass o example,
ege a ion and o es s), C_below (ca bon densi y con ained in belowg ound biomass;
hese a e plan and ee oo s), C_soil (ca bon densi y in soil) and C_dead (ca bon
densi y in decaying ma e ). They mus be w i en in he same way o else he model
will e u n an e o and ail o calcula e he ca bon s ock. The able should look as
shown below. The alues in his able we e consolida ed om he annex o a pape
assessing he impac o land co e changes in ecosys em se ices in Bo deaux (Cab al,
8
Fege , Le el, Chambolle, & Basque, 2016). Since he C_below, C_soil and C_dead
pools we e no being used, hey we e no included in he able.
lucode
LULC_name
C_abo e
1
A i icial su aces
1.13
2
Ag icul u al a eas
8.12
3
Fo es and semi na u al
a eas
19.78
4
We lands
1.17
5
Wa e bodies
0
Table 1: Ca bon pool alues o abo e g ound (Cab al e al., 2016)
3.1 De ailing he p og ession o LULC changes in Ge many om 1990-2018
This p ocess was ca ied ou in he change analysis ab o he LCM in Te Se model.
The cu en and u u e LULC maps we e en e ed in he model and hen i would
calcula e he a ea con e ing om one LULC ype o he o he . Addi ionally, he model
was able o calcula e he ne change aking place in each LULC ype, he gains and
losses occu ing and inally he in e ac ions be ween he LULC ypes. Tha is,
con ibu ion o change in one LULC ype by o he LULC ypes; how much a ea o
land in Hec a es a ce ain LULC ype was gaining o losing o he o he LULC ypes.
The aim o ca aloguing he LULC changes aking place o e he 28 yea s is o be able
o d aw ela ionships be ween he LULC changes and amoun o ca bon s ock in he
ege a ion.
3.2 Ca bon s ock end calcula ion
In o de o policymake s o adequa ely make decisions conce ning landscape
managemen , in o ma ion abou how much and whe e ca bon is s o ed is essen ial.
A e eclassi ica ion o he o iginal CLC maps, he esul ing maps we e clipped using
he Ge many adminis a i e bounda ies shape ile (“Download da a by coun y | DIVA-
GIS,” n.d.) in o de o be le wi h he s udy a ea. The Ge many LULC maps we e hen
loaded on o InVEST model wo a a ime, he cu en map which is he yea he model
begins he calcula ions om, and he u u e map which is he yea he model uses o
calcula e he ne change in ca bon seques a ion o e he se pe iod. The esul ing
alues can ei he be nega i e, which signi ies he loss o ca bon in o he a mosphe e
as ca bon dioxide, o posi i e, which will indica e ha ca bon was seques e ed. The
9
model wo ks by es ima ing he ne amoun o ca bon s o ed in he landscape o e he
du a ion o ime se by he wo land co e s. I app oxima es a alue by agg ega ing he
amoun o ca bon s o ed in he ca bon pools, (o in he case o his s udy, he
abo eg ound biomass) based on he land use maps and he new classes a e
eclassi ica ion (“Ca bon S o age and Seques a ion — InVEST 3.8.0 documen a ion,”
n.d.). Calcula ions done by he model a e ca ied ou pixel by pixel and he esul s a e
displayed as as e ou pu s o s o age, seques a ion and alue, as well as agg ega e
o als, measu ed in Megag ams o ca bon pe pixel.
The inal maps p oduced om his p ocess a e ca bon s ock o each o he i e yea s,
illus a ing he a ia ion in ca bon s ock h oughou he coun y. The o he maps
showed ca bon change om one yea o he o he , o example, he amoun o ca bon
s o ed in he landscape o los o he a mosphe e be ween 2000-2006. This aided in
showing he ca bon s ock change end obse ed o e he 28-yea pe iod. F om he
ca bon s ock changes i is hen possible o analyse how any LULC changes which may
ha e occu ed o e he yea s could ha e a ec ed he changes in amoun o ca bon
s ock: hus, p o ing o disp o ing he ac ha LULC changes ha e an e ec on ca bon
seques a ion.
3.3 P ojec ing u u e scena ios o ca bon seques a ion.
Scena ios enable policy make s o en ision he u u e in a con olled en i onmen
whe e hey a e also able o make necessa y adjus men s and co ec ions wi hou any
e ec on he eal wo ld. They a e able o do his using assump ions and ollowing
pa e ns o da a o p oduce isual ep esen a ions (Sha ma e al., 2018). Simple
scena ios only equi e in o ma ion abou he new policy o be implemen ed o a new
plan/p ojec unde conside a ion. Complex scena ios on he o he hand equi e
in o ma ion on d i e s o change on he landscape e.g., new policies, in as uc u al
de elopmen o clima e change da a. The scena ios modelled o his s udy a e
examples o complex scena ios (Ca e , La Ro e e, Jones, Leemans, & Nakiceno ic,
2001).
Ha ing analysed he p e ious yea s’ CLC maps and ca bon s ock/change maps, we
now need o come up wi h u u e land co e maps. The goal o his is o be able o
calcula e how much ca bon is likely o be s o ed in he landscape o los o he
10
a mosphe e in he u u e. Se e al app oaches ha e been used o model LULC change
including machine lea ning app oaches, cellula app oaches, economic app oaches,
agen -based app oaches and hyb id app oaches (Megahed, Cab al, Sil a, & Cae ano,
2015). This s udy ea u ed a hyb id app oach ha combined a cellula app oach which
is he CA Ma ko echnique in he Land Change Modelle (LCM) embedded in he
Te Se so wa e and a machine lea ning app oach using Mul i-Laye Pe cep on unde
he same so wa e (Yi saw, Wu, Shi, Temesgen, & Bekele, 2017). The me hodology
ha was ollowed in o de o ge he u u e LULC maps can be summa ised in Figu e
2 below.
Figu e 2: Land Change Modelle me hodology
The i s s ep in ol ed in his p ocess was ca ying ou change analysis o explo e he
in e ac ion wi hin he land uses be ween he yea s 1990-2018. This is whe e we a e
able o ell wha pe cen age o one land use was con e ed in o ano he land use o
which land uses pe sis ed, i.e., did no expe ience any change (Voigh , He nandez-
Aguila , Ga cia, & Gu ie ez, 2019). The model was ins uc ed o igno e ansi ions
unde 50,000 Ha in o de o be le wi h he mos signi ican ansi ions ha occu ed
in he 28-yea pe iod. The ollowing s ep is o model he ansi ion po en ials (Be g,

11
Roge s, & Mineau, 2016). The scena ios o answe he esea ch ques ion a e modelled
in his s ep. To do his, ou di e en scena ios we e modelled:
• In he i s one which was he baseline scena io, o he wise called Business as
Usual (BAU), no changes we e made. All he six ansi ion po en ials ha
esul ed om he change analysis s age we e used o p ojec he 2048 LULC
map wi h no es ic ions pu in place.
• The second scena io was he de elopmen scena io, in his case, a i icial
su aces we e no allowed o con e o any o he land use. This is because in
a case whe e a coun y p io i ises de elopmen , mo e and mo e land is needed
o allow o he cons uc ion o new in as uc u e. The e o e, ou di e en
ansi ion po en ials we e ou lined; ag icul u al a eas o a i icial su aces,
ag icul u al a eas o o es and semi na u al a eas, o es and semi na u al a eas
o a i icial su aces and inally, o es and semi na u al a eas o ag icul u al
su aces. These we e all modelled unde one sub-model.
• In he hi d scena io, which will be e e ed o as conse a ion 1 he ea e ,
h ee ansi ions we e modelled. These a e a i icial su aces o ag icul u al
a eas, a i icial su aces o o es and semi na u al a eas, and ag icul u al a eas
o o es and semi na u al a eas. In his scena io, o es and semi na u al a eas
we e es ic ed om unde going any con e sion since we a e ying o
maximise he amoun o ca bon being seques e ed in o he landscape.
• The inal scena io, conse a ion 2, is a modi ica ion o scena io 3. The poin
he e is o come up wi h a mo e achie able conse a ion scena io. The e o e,
he ansi ions modelled we e a i icial su aces o o es and semi na u al a eas,
a i icial su aces o ag icul u al a eas, ag icul u e o o es and semi na u al
a eas, and inally, ag icul u al a eas o a i icial su aces. This las ansi ion
was no included in he conse a ion 1 scena io.
I is impo an o no e ha he ou di e en scena ios we e modelled one by one, I.e.
he sub-models o ansi ion po en ial in each scena io we e un as one model, bu his
p ocess was done sepa a ely o each scena io. Following he modelling o he
scena ios, he independen a iables a e now added o he model. Fo his esea ch, he
independen a iables used we e;
12
• a Boolean cons ain map showing a eas whe e de elopmen was es ic ed
om aking place as 0 and he es as 1;
• a Digi al Ele a ion Model (DEM) o Ge many;
• a as e map o dis ance o oads in Ge many and;
• dis ance om Ge man u ban a eas, all o which we e go en om he same
sou ce (“Download da a by coun y | DIVA-GIS,” n.d.).
Independen a iables used du ing land change modelling a e ac o s ha a e likely o
in luence LULC changes. They we e all p epa ed in A cGIS using he Euclidean
dis ance ool. Be o e using he independen a iables, hey can be es ed o hei
po en ial explana o y powe using C ame ’s V es . This is a quan i a i e measu e o
associa ion on a scale o 0.0 which sugges s ha he e is no associa ion, o 1 which
sugges s pe ec co ela ion (Eas man, 2009). Howe e , his does no mean ha he
pe o mance o he model is au oma ically imp o ed. This is because he ela ionship
be ween he a iables and he model is complex and he ma hema ical equi emen s
canno be co obo a ed (Eas man, 2009).
The (LCM) embedded in Te Se has h ee modelling op ions namely, he Mul i-laye
Pe cep on (MLP) Ne wo k, Simila i y-Weigh ed Ins ance-based Machine Lea ning
(SimWeigh ) and Logis ic Reg ession. O hese, only he MLP allows modelling o
mo e han one ansi ion a he same ime. While bo h he MLP and SimWeigh
p oduce he bes esul s in ansi ion modelling (Te Se Tu o ial 1987, n.d.), he MLP
was p e e ed o his s udy due o i s abili y o un mo e han one ansi ion a he
same ime. T ansi ions should only be modelled oge he i he unde lying d i e s o
change a e belie ed o be simila .
A e modelling one scena io and inpu ing he independen a iables, he model is un
in o de o ge he ansi ion po en ial maps. These a e gene a ed by ex ac ing samples
om he LULC maps, p o ided in he beginning, ha wen h ough he ansi ions
being modelled, in addi ion o hose ha had he po en ial o unde go changes bu did
no . When his s ep is comple ed, he model agg ega es he esul s in h ml o ma and
p oduces a numbe o ansi ion po en ial maps equi alen o he numbe o ansi ions
ha we e ini ially modelled.
13
The ollowing s age is he change p edic ion ab. This is he las s ep o ge he u u e
LULC maps needed o his sec ion o he s udy. The amoun o change aking place
pe ansi ion can be modelled in wo ways. One is o use a Ma ko Chain analysis
embedded in he LCM module in Te Se , he o he is o speci y he ansi ion
p obabili y ma ix om an ex e nal sou ce. Fo his s udy, he Ma ko Chain analysis
was used. The ou pu o his s age is a ha d and a so p edic ion model, he la e being
op ional. The ha d p edic ion model is based on a compe i i e land alloca ion model
ha sha es cha ac e is ics wi h mul i-c i e ia decision p ocess. On he o he hand, he
so p edic ion model p oduces a map indica ing he p opensi y o change o he
speci ied se o ansi ions (Eas man, 2009). The in o ma ion used in his ab is
dependen upon wha was speci ied in he p e ious abs, speci ically he ansi ion
po en ial ab. The de aul me hod used o de e mine u u e occu ences o change is by
he use o a Ma ko Chain. This p ocess in ol es es ablishing he s a e o a sys em
based on i s p e ious condi ion and he p obabili y o change occu ing om one s a e
o he o he (Subedi, Subedi, & Thapa, 2013). The e o e, he model deduces he
p obabili y o a pixel changing om one land use ype o ano he . The Ma ko model
has been used in qui e a numbe o s udies in ecological modelling (Subedi e al., 2013;
Yi saw e al., 2017; Za andian e al., 2017). I is impo an o no e ha since he
objec i e he e is o model u u e LULC maps, he e is no way o de e mine how
accu a e hese maps would be because o he lack o an es ablished LULC map wi h
which o ca y ou a compa ison.
Following he comple ion o he change p edic ion s age, he ha d p edic ion maps
p oduced we e hen expo ed o A cGIS, adjus men s o he o ma made and he ea e
hey we e loaded on o he InVEST model. He e, he p ocess ca ied ou o calcula e
ca bon seques e ed in he landscape in he pas 28 yea s was epea ed. The only
di e ence is ha in his case 2018 was used as he cu en yea map and he p edic ed
land co e maps o 2048 used as he u u e yea .
14
4. Resul s
4.1 LULC changes in Ge many om 1990-2018
Acco ding o Figu e 3 below, du ing he ime pe iod be ween 1990 and 2000, A i icial
su aces saw he mos gain while ag icul u al a eas aced he mos loss. The land los
om ag icul u al a eas ansi ioned o a i icial su aces. Fo es and semi na u al a eas
oo gained some land om ag icul u al a eas, and he e was mo e land gained han was
los in his class. F om he yea 2000 o 2006 ag icul u al a eas and a i icial su aces
a e s ill expe iencing he mos change. A i icial su aces gained 0.41% o he o al
land a ea and los 0.1% while ag icul u al a eas gained 0.09% and los 0.47%. Du ing
his pe iod o es and semi na u al a eas los mo e land, 0.11% han in he p e ious
1990-2000 pe iod.
Figu e 3: Gains and losses end o 28 yea s
Be ween 2006 and 2012 he land gained by he a i icial su aces does no seem o be
as signi ican as wha occu ed be ween 1990 o 2006. On he o he hand, ag icul u al
a eas appea o ha e gained e en mo e land han was gained be ween 1990-2006. In
o es and semi na u al a eas he same seems o be he case, wi h a sligh inc ease in
land gained o e land los . Be ween he yea 2012-2018, ag icul u al a eas ha e
con inued o gain mo e land 1.52% while losing almos jus as much, 1.43%. Gains
and losses in o es and semi na u al a eas seem o ha e emained p e y much he
same a 1.02% and 1.12% espec i ely, in compa ison o 2006- 2012. The e was also
21
5. Discussion
Be o e e iewing he esul s o he s udy and p o iding an analysis, i is essen ial o
poin ou ha he models used while use ul and ela i ely easy o use, had a ew
limi a ions, as all models do. The InVEST model o example, assumes ha he land
use classes a e no gaining o losing ca bon o e ime wi hin he classes hemsel es.
This is no he case, especially when i comes o o es s and semi na u al a eas. The
amoun o ca bon s o ed in o es s educes wi h hei age. As hey become mo e
ma u e, he ca bon hey s o e le els ou (Zhu, Song, & Qin, 2019), and his canno
cu en ly be cap u ed by he model. Due o his assump ion, he only changes
conside ed by he model a e hose ha occu as a esul o changes om LULC ype
o ano he (“Ca bon S o age and Seques a ion — InVEST 3.8.0 documen a ion,”
n.d.). The LCM in Te Se model on he o he hand only deals wi h physical ac o s.
Socio-economic and demog aphic ac o s canno be added o he model, his means
ha change can only be modelled wi h a iables such as dis ance o oads and u ban
cen es, ele a ion, slope and no popula ion o economic da a, bo h o which ha e been
known o a ec change in LULC. In addi ion, i is no possible o specialise he
pa ame e s e.g., speci ying by how much pe cen age one would like a pa icula land
co e o change when modelling he scena ios. The e o e, one is s uck only modelling
he p obabili y o he pixels in one LULC ype changing o ano he , bu no by wha
pe cen age. This pu s a limi on how much change can be modelled. Ne e heless, he
models p o ided in o ma ion ha will aid in answe ing he esea ch ques ions b ough
up by his s udy.
5.1 LULC changes and ca bon s ock in Ge many om 1990-2018
The decision on which o he ou majo ca bon pools speci ied in he IPCC epo o
use o his s udy was based on he esea ch design and a ailabili y o da a. Since he
ocus o he esea ch is o explo e he ela ionship be ween LULC and ca bon
seques a ion, i makes sense o use abo eg ound biomass. Adding he o he h ee
ca bon pools would ha e p obably yielded be e esul s in e ms o he amoun o
ca bon s ock (Un ccc, 2015). Howe e , i would ha e p o en di icul a e wa ds o
know exac ly wha pe cen age o he ca bon s ock was as a di ec esul o LULC
changes wi h he ools a ailable a he s udy’s disposal.

22
While calcula ing ca bon seques a ion om 1990-2018, he changes obse ed om
one yea o he o he we e almos negligible in compa ison o he o al amoun o
ca bon s ock. In 1990, 393,871,051.9 Mg o C was s o ed in abo eg ound ege a ion.
By 2018 he o al amoun had inc eased o 394,471,457.7 Mg o C (11.04 MgC pe
Ha) he e o e indica ing an inc ease o only 0.15%. This is suppo ed by he ac ha
he changes occu ing in o es and semi na u al a eas we e e y small. Majo i y o he
changes we e occu ing be ween ag icul u al and a i icial su aces. Howe e , in one
ime pe iod, be ween 2006-2012, he e was a no able inc ease in he amoun o ca bon
s ock as seen in Figu e 8. A he same ime, ag icul u al a eas gained e en mo e land
han in he p e ious ime pe iod, 2000-2006 as seen in Figu e 3.
This could mean ha he e is mo e po en ial o ca bon seques a ion in he ag icul u e
sec o ha needs o be exploi ed. This can be used in conjunc ion wi h measu es
applied o o es and semi na u al a eas in o de o ensu e ha all he a enues o
ca bon seques a ion a e maximised. Acco ding o (Un ccc, 2008), ha mony be ween
ag icul u e- ela ed clima e change policies and sus ainable de elopmen is necessa y
i he e is hope o en ice a me s, policymake s and land manage s o adop ag icul u al
mi iga ing p ac ices.
5.2 Modelling o u u e scena ios
I is impo an o no e ha scena ios a e “nei he p edic ions no o ecas s” (IPPC,
2014). Ins ead, hey pain a pic u e o p obabili ies, e en s ha a e likely o happen i
ce ain a iables a e applied. I was challenging o model scena ios ollowing he IPCC
guidelines on scena ios as emissions om land use change ha e no been as well
documen ed as hei coun e pa s, ene gy- ela ed emissions. The e o e, he scena ios
modelled in his s udy ollowed a simple pa h o ;
a) Business as Usual, i.e., wha would happen i no o m o in e en ion
happened. The e was an inc ease in ca bon s ock by 0.22%. In his scena io,
LULC is allowed o de elop wi hou any in e e ence o es ic ions. This
p o ides policymake s wi h a glimpse o wha a eas could use some
imp o emen i hey hope o ap in o he abili y o he LULUCF sec o o aid
in clima e change mi iga ion.
23
b) De elopmen , wha is likely o happen i a coun y ocuses on de elopmen
and ails o p io i ise conse a ion. This is a s a e in which he go e nmen
seeks o maximise economic g ow h by expanding land uses which can be
exploi ed o example, mo e indus ializa ion, mo e oads, mo e ag icul u e.
A e he modelling and calcula ion o u u e ca bon s ock, a dec ease in ca bon
s ock le els by 0.96% om 2018 was obse ed. Since his scena io depic s
wha is likely o happen i a coun y decided o p io i ise de elopmen o e
conse a ion, mo e land is needed o ensu e mo e de elopmen . In mos cases
his land ends up being aken om he o es sec o . This is a scena io ha is
no likely o happen in Ge many seeing as i is al eady a de eloped coun y
he e o e nega ing he need o p io i ise de elopmen . I was howe e modelled
in o de o se e as a cau ion by illus a ing wha could happen should a coun y
decide o go in ha di ec ion.
c) Conse a ion. Wha kind o esul s would we expec i conse a ion was
conside ed a op p io i y? A ound 430 million onnes o a mosphe ic ca bon
dioxide is seques e ed in o es co e in Eu ope (Schelhaas e al., 2015). I
he e o e s ands o eason ha unde conse a ion scena ios he aim would be
p ese ing he exis ing o es co e as much as possible. The conse a ion
scena ios we e spli in o wo o allow o di e en iewpoin s.
i. Conse a ion 1: The i s one was a scena io whe eby no new a i icial
su aces we e modelled. Ge many has been wo king on educing he
amoun o land ha is being pa ed o e and hey had hoped o educe
his numbe o 30 Ha pe day by 2020 (“Land use educ ion |
Umwel bundesam ,” n.d.) wi h he objec i e o slowing down land use.
This allows o na u al a eas o emain as hey a e o a he e y leas
o no unde go cemen a ion which e ec i ely cu bs any o m o
seques a ion ha would occu o he wise. This scena io saw he highes
inc ease in ca bon s ock by 4.16% om 2018 le els. I would be ideal
and could be possible especially since Ge many’s popula ion is
o ecas ed o be educe o 80 million (“Ge many Popula ion (2020) -
Wo ldome e ,” n.d.). Which ideally would mean ha he e would be
24
less need o expansion o ag icul u al land and he e o e abandoned
pa cels could be used o a o es a ion.
ii. Conse a ion 2: The second conse a ion scena io had he same 3
ansi ions modelled in conse a ion 1 bu wi h an addi ional ansi ion
whe eby ag icul u al a eas could po en ially be con e ed o a i icial
su aces. This saw an inc ease o 0.41% o ca bon s ock om 2018
le els. This is sligh ly mo e han wha would be seques e ed unde he
BAU scena io, bu i is ba ely enough o make a signi ican di e ence.
Howe e , his is he mo e p obable scena io be ween he wo
conse a ion scena ios. E en wi h a p ojec ed dec ease in popula ion i
is highly likely ha a i icial su aces will con inue o g ow as mo e
people con inue o mo e o he u ban cen es he eby inc easing he
need o mo e a i icial su aces. This would in u n hen mean ha
he e is less land a ailable o g ow mo e o es s.
O he ou scena ios modelled, h ee o hem ea u e an inc ease in ca bon s ock, his
is mainly d i en by he inc ease in o es co e . This is e idenced by he ac ha he
conse a ion 1 scena io, which had he mos a ea o land being occupied by o es and
semi na u al a eas, had he highes inc ease in ca bon s ock a 4.16% wi h a a e o
547,387.95 Mg o C (0.153 MgC pe Ha) pe yea om 2018 le els. The amoun o
ca bon s ock om 1990 o 2048 is summa ised in Figu e 12 below.
25
Figu e 12: Ca bon s ock summa y om 1990-2048
As epo ed by (Bondeau e al., 2007), he u u e o e es ial ca bon balance in Eu ope
is qui e unclea . Despi e i being known ha conse a ion and eplenishing o o es s
ansla es o an inc ease in ca bon up ake, i is no known whe he his end is se o
con inue o be he case in he u u e. Wi h he modelled scena ios, policymake s a e
able o ge a glimpse o wha can be expec ed i ce ain pa hs a e aken. As a esul ,
hey can in e ene and add mo e s ingen measu es o modi y he exis ing ones i hey
hope o cu b clima e change be o e i is oo la e.
A e explo ing all hese di e en scena ios, i is e iden ha in o de o secu e
Ge many’s abili y o seques e mo e ca bon in he LULUCF sec o , mo e policies ha
a ge he conse a ion and eplenishmen o o es co e need o be implemen ed.
Howe e , i p o ec i e measu es a e applied o o es s wi hou examining d i e s o
change and sugges ing ways o add essing hem, hese policies would only educe a
small pe cen age o o al emissions om he LULUCF sec o . Thus e ec i ely leading
o “c oss-biome leakage” which will se iously hampe any e o s o mi iga e agains
clima e change (S assbu g, La awiec, C eed, & Nguyen, 2014). As a esul , his calls
o mo e measu es o be applied in he ag icul u e sec o as well. Figu e 14 below was
consolida ed by con e ing he ca bon s ock alues go en om he ou modelled
scena ios, which a e measu ed in Mg o C (Megag ams o Ca bon) o Kilo onnes o
26
CO2. The con e sion o 1 giga onne o C = 3.664 billion onnes o CO2 was used. This
was done in o de o acili a e compa ison be ween how much CO2 pe yea would be
sunk in each scena io in ela ion o he o al emissions o Ge many’s las in en o y
yea which was 2017.
Figu e 13: To al Ge many CO2 emissions measu ed in Figu e 14: Es ima ed ca bon s ock a e pe yea
Kilo onnes o CO2 (“G eenhouse Gas In en o y Da a - Time Se ies - Annex I,” n.d.)
In 2016, almos 58 million onnes o CO2 equi alen s (58,000 k o CO2) was s o ed
in he LULUCF sec o in Ge many (Minis y e al., 2018). 96.5% o his igu e was
s o ed in o es s. This poin s o he impo ance o his sec o in clima e change
mi iga ion. Howe e , om he Figu e 14 abo e, we see ha in he bes -case scena io,
conse a ion 1, a ound 50,000 k CO2 equi alen s is s o ed in he landscape pe yea .
When his alue is compa ed o he emissions pe yea , Figu e 13, in which 2017 saw
a ound 800,000 k C02 equi alen s emi ed, i ba ely makes a den . Ge many’s goal is
o ha e he ag icul u e sec o emi ing no mo e han 61 million onnes (61,000 k o
CO2). Going a his a e, Ge many is likely o ail o mee i s a ge s again come 2030.
Consequen ly, i he e is any chance o sal aging he si ua ion, he bes op ion would
be o combine a i icial means wi h na u al means o ca bon s o age in o de o be able
o achie e ca bon neu ali y in Ge many (Hood, 2007).

27
Figu e 15: The ou illus a i e model pa hways sugges ed o manage ca bon s o age and seques a ion (Hood,
2007).
F om Figu e 15 abo e, ei he scena io P2 o P3 would aid in achie ing his. Bo h
scena ios combine na u al and a i icial me hods o ca bon seques a ion whe eas in
P1 he e is no a i icial in e en ion and in P4 he e is a hea y eliance on echnology
o ca bon cap u e. Addi ionally, he go e nmen o Ge many plans o exploi he
abili y o soil o ac as a ca bon sink by suppo ing a ge ed plan ing o ca bon
abso bing plan species and encou aging be e p o ec ion and es o a ion o moo and
we lands (“Ge many’s Clima e Ac ion P og amme 2030 | Clean Ene gy Wi e,” n.d.).
28
6. Conclusion
Emissions educ ion ia implemen a ion o Reducing Emissions om De o es a ion
and Fo es Deg ada ion (REDD+) wo ks e y well o clima e change mi iga ion in
opical de eloping coun ies. Howe e , de eloped coun ies would bene i mo e om
e o es a ion, o es managemen and ha es ed wood p oduc s (FAO, 2016). This
makes sense in he case o Ge many especially because, he coun y being de eloped
means ha he e is no a lo o land a ailable o plan ing mo e o es s. E idence o
his can also be seen om he LULC maps. The e o e, he mi iga ing po en ial o
Ge many’s o es s needs o be ully exploi ed in o de o ha e hem sinking mo e
ca bon ha hey emi . This can be done by adop ing low-ca bon in ensi y echnology
and en o cing p ope managemen o he exis ing o es s he eby esul ing in he use
o o es esou ces wi hou inc easing emissions om hem (FAO, 2016).
This he e o e means ha i Ge many hopes o maximise on he abili y o he LULUCF
sec o o aid in clima e change mi iga ion, mo e esea ch needs o be done in o de o
ind las ing solu ions ha would encou age mo e ca bon seques a ion. Howe e e en
wi h his in mind, o es s a e no immune o clima e change impac s o example,
s o ms, human in luences and na u al dis u bances. The e o e, he measu es pu in
place should conside his and make he necessa y accommoda ions. Any u he
esea ch conduc ed in his a ea should conside inco po a ing s akeholde s’ inpu in
he scena io modelling p ocess as ha would ha e p o ed o be o g ea bene i o he
s udy. In addi ion, inclusion o he o he h ee ca bon pools i.e., belowg ound ca bon,
soil o ganic ca bon and dead ma e ca bon would ce ainly g ea ly imp o e he esul s.
29
Bibliog aphic Re e ences
Annual Repo 2018–2019 | Wo ld Economic Fo um. (n.d.). Re ie ed Feb ua y 17,
2020, om www.we o um.o g/ epo s/annual- epo -2018-2019
Ansa i, A., & Golabi, M. H. (2019). P edic ion o spa ial land use changes based on
LCM in a GIS en i onmen o Dese We lands – A case s udy: Meighan
We land, I an. In e na ional Soil and Wa e Conse a ion Resea ch, 7(1), 64–
70. h ps://doi.o g/10.1016/j.iswc .2018.10.001
Bellassen, V., & Luyssae , S. (2014). Ca bon seques a ion: managing o es s in
unce ain imes. Na u e, 506(7487), 153–155. h ps://doi.o g/10.1038/506153a
Be g, C., Roge s, S., & Mineau, M. (2016). Building scena ios o ecosys em
se ices ools: De eloping a me hodology o e icien engagemen wi h expe
s akeholde s. Fu u es, 81, 68–80. h ps://doi.o g/10.1016/j. u u es.2015.10.014
Bondeau, A., Ca e , T. R., C ame , W., E ha d, M., P en ice, I. C., Regins e , I., …
Sykes, M. (2007). P ojec ed Changes in Te es ial Ca bon S o age in Eu ope
unde Clima e and Land-use Change ,. 380–401.
h ps://doi.o g/10.1007/s10021-007-9028-9
Cab al, P., Fege , C., Le el, H., Chambolle, M., & Basque, D. (2016). Assessing he
impac o land-co e changes on ecosys em se ices: A i s s ep owa d
in eg a i e planning in Bo deaux, F ance. Ecosys em Se ices, 22, 318–327.
h ps://doi.o g/10.1016/j.ecose .2016.08.005
Ca bon S o age and Seques a ion — InVEST 3.8.0 documen a ion. (n.d.). Re ie ed
Feb ua y 14, 2020, om h p:// eleases.na u alcapi alp ojec .o g/in es -
use guide/la es /ca bons o age.h ml
Ca e , T., La Ro e e, E., Jones, R., Leemans, R., & Nakiceno ic, N. (2001).
De eloping and applying scena ios. Wo king G , 145–190.
Ciais, P., Sabine, C., Bala, G., Bopp, L., B o kin, V., Canadell, J., … Tho n on, P.
(2013). The physical science basis. Con ibu ion o wo king g oup I o he i h
assessmen epo o he in e go e nmen al panel on clima e change. Change,
IPCC Clima e, 465–570. h ps://doi.o g/10.1017/CBO9781107415324.015
30
CORINE Land Co e — Cope nicus Land Moni o ing Se ice. (n.d.). Re ie ed
Feb ua y 3, 2020, om land.cope nicus.eu/pan-eu opean/co ine-land-co e
Download da a by coun y | DIVA-GIS. (n.d.). Re ie ed Feb ua y 8, 2020, om
www.di a-gis.o g/gda a
Eas man, J. (2009). IDRISI Taiga: Guide o GIS and Image P ocessing Volume -
Manual e sion 16.02. (Augus ), 325.
Eu ope :: Ge many — The Wo ld Fac book - Cen al In elligence Agency. (n.d.).
Re ie ed Feb ua y 17, 2020, om www.cia.go /lib a y/publica ions/ he-wo ld-
ac book/geos/gm.h ml
FAO. (2016). Fo es y o a low-ca bon u u e. FAO Fo es y Pape 177.
Fo es s and Ag icul u e. (n.d.). Re ie ed Feb ua y 14, 2020, om
h ps://ec.eu opa.eu/clima/policies/ o es s_en
Ge many’s Clima e Ac ion P og amme 2030 | Clean Ene gy Wi e. (n.d.). Re ie ed
Feb ua y 14, 2020, om h ps://www.cleanene gywi e.o g/ ac shee s/ge manys-
clima e-ac ion-p og amme-2030
Ge many’s g eenhouse gas emissions and clima e a ge s | Clean Ene gy Wi e. (n.d.).
Re ie ed Feb ua y 14, 2020, om
h ps://www.cleanene gywi e.o g/ ac shee s/ge manys-g eenhouse-gas-
emissions-and-clima e- a ge s
Ge many Popula ion (2020) - Wo ldome e . (n.d.). Re ie ed Feb ua y 7, 2020, om
www.wo ldome e s.in o/wo ld-popula ion/ge many-popula ion/
G eenhouse Gas In en o y Da a - Time Se ies - Annex I. (n.d.). Re ie ed Feb ua y
7, 2020, om di.un ccc.in / ime_se ies
Hood, R. (2007). Global Wa ming. A Companion o Applied E hics, 674–684.
h ps://doi.o g/10.1002/9780470996621.ch50
Indica o : G eenhouse gas emissions | Umwel bundesam . (n.d.). Re ie ed Feb ua y
14, 2020, om h ps://www.umwel bundesam .de/en/indica o -g eenhouse-gas-
emissions#a -a-glance