ARTEMIS Ve sion 1.0
A Reac i e T anspo Enhanced-wea he ing Model In Soils
Use Guide
Lyla L. Taylo
22 No embe 2025
Le e hulme Cen e o Clima e Change Mi iga ion*
Di ec o : Da id J. Bee ling
School o Biosciences
Uni e si y o She ield
*Funded by he Le e hulme T us (Le e hulme Resea ch Cen e g an numbe RC-2015-029 o Da id
J. Bee ling)
Con en s
1 In oduc ion 1
1.1 Ins alla ion on Linux sys ems ............................... 1
1.2 De elopmen his o y .................................... 2
1.3 Two geochemical models .................................. 4
1.4 Disclaime .......................................... 4
2 Inpu iles 5
2.1 Rock-MIP pa ame e iles ................................. 5
2.2 Rock-MIP ini ialChemis y ile ............................... 5
2.3 Phase da a ile ....................................... 7
2.4 Ini ial soil laye da a ile .................................. 9
2.5 Feeds ock PSD ile ..................................... 13
2.6 Time able o e en s ..................................... 14
2.7 Timese ies inpu iles .................................... 14
2.8 Dep hwise ime se ies inpu iles .............................. 15
2.9 Obse a ions iles ...................................... 16
2.10 Files in subdi ec o y “ph eeqcdbinclude iles” ....................... 17
3 Model pa ame e s and se ings 19
3.1 The op ional de aul se ings.m sc ip ........................... 19
3.1.1 Su ace a ea me hod ................................ 19
3.1.2 E apo anspi a ion me hod ............................ 23
3.1.3 EXCHANGE me hod ................................ 23
3.1.4 MIX di ec ion .................................... 24
3.1.5 Soil gas phase op ions ............................... 24
4 Ene gy Fa m example 27
4.1 Inpu iles and sc ip s .................................... 27
i
ii
4.2 How o do an example Ene gy Fa m un ......................... 28
4.3 Loading and displaying he Ene gy Fa m esul s ..................... 32
4.4 Ene gy Fa m calib a ion uns ............................... 32
4.4.1 Phospho us dynamics du ing calib a ion uns .................. 34
5 Na iga ing he so wa e 37
5.1 The p ep ocesso : se upda as uc.m ............................ 37
5.1.1 Files o be w i en p io o he i s model un .................. 38
5.1.2 checkphase ype.m .................................. 39
5.1.3 implici calc.m .................................... 40
5.2 PHREEQC w appe : ge ph eeqc un.m .......................... 41
5.3 Da a isualisa ion u ili ies ................................. 42
Bibliog aphy 45
Lis o Figu es
1.1 O e all s eps o unning his model ............................ 3
4.1 Ene gy Fa m calib a ion uns ............................... 34
4.2 Topsoil P pools du ing calib a ing uns .......................... 35
4.3 P up ake du ing calib a ion uns ............................. 36
5.1 P ep ocesso : unc ions called ............................... 39
5.2 PHREEQC w appe : unc ions called ........................... 42
5.3 PHREEQC inpu ile o he i s imes ep ........................ 43
5.4 PHREEQC inpu ile o subsequen imes eps ...................... 44
iii
Lis o Tables
1.1 So wa e equi ed o un ARTEMIS ............................ 3
2.1 Inpu .cs iles o he p ep ocesso ............................ 6
2.2 Time a iables allowed o imese ies and ime able da a ................ 6
2.3 Rock-MIP pa ame e iles: example da a ......................... 7
2.4 Inpu da a o all solid ino ganic phases o be modelled ................. 8
2.5 Allowed designa ions o kine ic eac ion o de s in phase able ............. 10
2.6 Soil laye ile column headings ............................... 10
2.7 Soil laye inpu pa ame e s ................................. 11
2.8 Time able s uc u e a ays ................................. 15
2.9 Time se ies iles o o cing da a .............................. 16
2.10 The ph eeqcdbinclude iles di ec o y ............................ 18
3.1 Feeds ock applica ions and ex ended uns ......................... 20
3.2 Pa ame e s con olling eac i e su ace a ea (RSA). ................... 20
3.3 P ep ocesso pa ame e s ela ed o so p ion ....................... 21
3.4 Addi ional pa ame e s used in his model ......................... 22
3.5 Op ions o se ing he SAme hod keywo d ........................ 23
3.6 Op ions o se ing he ETme hod keywo d ........................ 24
3.7 Op ions o he soilco2op keywo d ............................ 25
3.8 O he keywo ds o soil gases ............................... 26
4.1 Ene gy Fa m calib a ion sc ip s and se ings ....................... 27
4.2 Fi e-yea Ene gy Fa m uns ................................ 28
4.3 MATLAB sc ip s o Ene gy Fa m uns h ough 2070 .................. 28
4.4 Example ou pu iles .................................... 30
4.5 Da a iles o Ene gy Fa m uns .............................. 31
4.6 Mine alogy o he blue idge me abasal .......................... 33
i
Chap e 1
In oduc ion
ARTEMIS is a de e minis ic, p ocess-based eac i e anspo model (RTM) o calcula ing ca bon
dioxide emo al (CDR) ollowing enhanced ock wea he ing (ERW) in soils. I aims o accoun o
as many soil p ocesses as possible, as ealis ically as possible, allowing use s o in es iga e p ocess
e ec s by u ning indi idual p ocesses on o o . These aims a e, o cou se, only pa ially ul illed, and
he e a e pi alls associa ed wi h his le el o complexi y. Fo example, use s should be wa y o ying
o calib a e e e y p ocess a a pa icula ield si e, gi en he conside able he e ogenei y o managed
soils. Ne e heless, his code may p o e use ul o he CDR, ERW and geochemical communi ies.
This use guide p o ides ins uc ions o ins alling ARTEMIS (Sec ion 1.1). I also explains he
s uc u e and pa ame e iza ions o ARTEMIS Ve sion 1.0, along wi h some examples showing how o
un he model. I assumes some expe ience wi h bo h MATLAB and wi h he unde lying geochemical
specia ion code (PHREEQC, Pa khu s and Appelo, 2013).
A p ep ocesso and PHREEQC w appe comp ise he bulk o he MATLAB code desc ibed he e
(Figu e 1.1). The s eps shown in Figu e 1.1 will be desc ibed in subsequen chap e s. Se e al
publicly-a ailable MATLAB so wa e packages a e also equi ed o un ARTEMIS (Table 1.1).
The code uns unde he Linux ope a ing sys em and has been success ully es ed on wo high-
pe o mance compu e s (HPCs) wi h di e en ecen e sions o MATLAB (Sec ion 1.2). The e is
no gua an ee wha soe e ha ARTEMIS will un on any o he sys em; I am no a Windows o Mac
use , so ha e ne e ied o ins all o un i on hose ope a ing sys ems.
1.1 Ins alla ion on Linux sys ems
No e: hese ins uc ions assume basic knowledge o bo h Linux and MATLAB, and ha MATLAB
is al eady ins alled.
To ins all ARTEMIS, download he comp essed a ball and unpack i wi h a xz :
a xz ARTEMIS_GMD_code_ 1.0_Sep-2025. a .gz
A di ec o y “soilmodel” will be c ea ed con aining all he m- iles (MATLAB code). The e a e se e al
subdi ec o ies:
1
2
ph eeqcdbinclude iles con ains he ARTEMIS PHREEQC da abases along wi h some addi ional
.cs iles (Sec ion 2.10).
examples con ains he da a and sc ip s o he Ene gy Fa m example uns (see Chap e 4).
The e is no Make ile, bu he e a e wo en i onmen a iables o he home di ec o y and a sc a ch
di ec o y in his code which should be de ined: $HOME and $SCRATCH ( he de aul di ec o y o
model ou pu iles). Fo example, I ha e de ined $SCRATCH in my .bash c ile:
expo SCRATCH=/mn /pa sc a ch/use s/$USER
Finally, he MATLAB s a up.m ile needs o add he main model di ec o y o he MATLAB pa h,
along wi h he di ec o ies con aining he MATLAB code o S oichTools and he SkillMe ics oolbox
(Table 1.1).
1.2 De elopmen his o y
De elopmen o ARTEMIS commenced in 2020 wi h he aim o modelling he e ec o ock dus
ea men s in a soy-maize-maize ield expe imen un by he Le e hulme Cen e o Clima e Change
Mi iga ion (LC3M) (Kan ola e al., 2017; Kan ola e al., 2023; Bee ling e al., 2024). The in e ace
wi h PHREEQC (Pa khu s and Appelo, 2013) was loosely based on MATLAB so wa e de eloped o
model CO2consump ion o he Hubba d B ook wollas oni e ea men (Taylo e al., 2021) bu was
also in o med by he PHREEQC sc ip w i en by Pe e Wade o he mesocosm s udy o Kelland
e al. (2020).
ARTEMIS has always included mo e complex biologically-media ed p ocesses such as nu ien up-
ake, decomposi ion, ni ogen cycling and soil pCO2based on inpu soil espi a ion. I is pa ly based
on es ablished land and soil models (Nei sch e al., 2011; Law ence e al., 23 Ma ch 2020) and on
ea lie geochemical models o he long- e m global ca bon cycle ha I de eloped a he Uni e si y
o She ield (Taylo e al., 2016; Taylo e al., 2012). Subsequen be a- es o p o o ype e sions
o ARTEMIS we e adap ed wi h he aim o modelling se e al o he LC3M expe imen s a bo h
labo a o y and ield scale, and o p oduce uns o a model in e compa ison p ojec o enhanced
wea he ing (Rock-MIP, Taylo e al., 2023; Taylo e al., 2024). Ve sion 1.0 dis ils code om se e al
o hese expe imen -speci ic p o o ypes, so he e is no gua an ee ha use s will ind ARTEMIS easy
o ollow o modi y.
ARTEMIS was de eloped on a succession o high-pe o mance compu e s (HPCs) a he Uni e si y
o She ield. The Ene gy Fa m example simula ions (see Chap e 4) we e un in ba ch mode on a
Dell Powe Edge C6420 unning SUSE Libe y Linux 7 wi h he slu m job scheduling sys em and
MATLAB e sion R2022a. Tha HPC was decommissioned almos immedia ely a e comple ing
hose example uns, bu be o e ARTEMIS was o be published. The code was ans e ed o i s
cu en home, a Dell Powe Edge R650 also unning SUSE Libe y Linux 7 wi h slu m and MATLAB
e sion R2023b. The “bes ” calib a ed example un o he Ene gy Fa m (see Table 4.1) was es ed
on he new HPC o check ha he ou pu s had no changed.
ARTEMIS does no equi e pa allel p ocessing o mul iple co es. I ha e occasionally done sho uns
on he MATLAB command line on my desk op compu e unning a ious e sions o Ubun u and
3
MATLAB, bu his is no he mos con enien way o unde ake a la ge s udy. In ba ch mode on
he HPC, he long Ene gy Fa m uns co e ing he pe iod 2016–2070 (Sec ion 4.1 and Table 4.3)
and ea ing apa i e as an “implici ” phase (see Sec ion 4.2 and he GMD manusc ip ) each ook
≃7hou s o un, while he i e-yea un wi h kine ic apa i e and he c ode implici sol e equi ed
3–4 hou s. The un wi h kine ic apa i e and he Runge-Ku a sol e an o o e 2.5 hou s be o e
c ashing a e 1270 ou o 1829 days. The quickes un imes we e 21–23 minu es o he i e-yea
uns wi hou so p ion; mos o emaining i e-yea uns equi ed 33–60 minu es. The e o e, he
abili y o un simul aneous jobs in he backg ound is use ul o ARTEMIS.
Figu e 1.1: O e all s eps o unning his model. See Chap e 4 o examples o how o un he so wa e.
Table 1.1: So wa e equi ed o un ARTEMIS
So wa e Requi ed? Language Pu pose
MATLAB equi ed Running he p ep ocesso and PHREEQC w appe
h ps://uk.ma hwo ks.com/p oduc s/ma lab.h ml
PHREEQC equi ed C and C++ Geochemical calcula ions
h ps://www.usgs.go /so wa e/ph eeqc- e sion-3
S oichTools equi ed MATLAB chemical no a ion pa sing
h ps://uk.ma hwo ks.com/ma labcen al/ ileexchange/29774-s oichiome y- ools
LC3M PSD
so wa e
op ional MATLAB Pa icle size dis ibu ion acking, de eloped by Ma k
Lomas (Bee ling e al., 2020)
No ye publicly a ailable. Con ac Da id Bee ling o mo e in o ma ion.
4
1.3 Two geochemical models
ARTEMIS is he second RTM o be published by he LC3M. Se e al la ge -scale LC3M publica ions
(Bee ling e al., 2020; Kan zas e al., 2022; Bee ling e al., 2025) employed a di e en RTM de eloped
independen ly by ma hema ician Ma k Lomas unde he auspices o he la e geochemis S e en A.
Banwa . Ma k’s code (Bee ling e al., 2025, h ps://doi.o g/10.5281/zenodo.10940280) calcula es
po en ial CDR based on eeds ock wea he ing, and is a mo e amenable o la ge-scale s udies and
emula ion han ARTEMIS as i equi es ewe inpu s and sho e un imes. Al hough I did no w i e
he code o ha model, I p o ided ad ice abou how o code he alkalini y calcula ions (Bee ling
e al., 2020), and de eloped equa ions desc ibing he o e all e ec o ni ogen cycling on pH (Kan zas
e al., 2022) and, mo e ecen ly, ca ion exchange (cu en ly unde de elopmen ).
1.4 Disclaime
This documen a ion, along wi h he code and ou pu iles o he example model uns, is in ended o
accompany a model desc ip ion pape in ended o submission o Geoscien i ic Model De elopmen
(GMD). I is eely a ailable on a “wa s and all” basis, wi h no gua an ees o any kind and no
echnical suppo . I will howe e y o answe ques ions ela ing o he ARTEMIS code and he
GMD manusc ip .
I would like o hank my coau ho s on he GMD manusc ip desc ibing ARTEMIS Ve sion 1.0 and
indeed o he pas and p esen membe s o he LC3M eam who p o ided many help ul sugges ions,
da a and suppo o e he yea s. The name ARTEMIS was chosen by he LC3M di ec o and
manusc ip coau ho , Da id Bee ling; i is a be e name han any I had come up wi h. Howe e ,
nei he he no any o he coau ho s o colleagues a e esponsible o he e o s, omissions o o he
p oblems wi h he Ve sion 1.0 code o documen a ion ha will undoub edly come o ligh . I am
en i ely o blame o hose.
11
Table 2.7 (see page 11) shows he equi ed pa ame e s ha ough o appea in his able, along
wi h some op ional ways o en e ing some o he da a. Fo example, he model equi es olume ic
po osi y, ield capaci y and wil ing poin , as well as he ini ial wa e con en . These may be speci ied
o each laye , o hey may be calcula ed om he soil ex u e.
As hese pa ame e s may no all ha e been measu ed, he p ep ocesso can calcula e hem using
he equi ed soil ex u e da a and one o he ollowing pedo ans e unc ions ( om he Communi y
Land Model Ve sion 5 as ep esen ed by Jinyun Tang in main/FuncPedo ans e Mod.F90):
Cosby5 Cosby e al. (1984, hei Table 5), de aul o his p ep ocesso and o CLM5
Cosby4 Cosby e al. (1984, hei Table 4)
NoilhanLaca e e1995 Noilhan and Laca è e (1995)
The desi ed pedo ans e unc ion may be speci ied in de aul se ings.m (Sec ion 3.1); o he wise
Cosby5 will be used.
By de aul , i no ini ial soil mois u e is gi en, he model s a s a ield capaci y o all cells.
Da a ela ed o he ca ion exchange capaci y a e equi ed, bu he e a e se e al di e en op ions:
P o ide CEC and exchangeable ca ion da a I he exchangeable ca ion da a do no add o
100% o he gi en CEC, exchangeable acidi y will be added (de aul : as p o ons).
P o ide CEC only Wi hou exchangeable ca ion da a, ARTEMIS will equilib a e he PHREEQC
EXCHANGE block wi h he solu ion.
P o ide exchangeable ca ion da a only CEC will be es ima ed as he sum o he da a p o ided.
Solid mine al o amo phous soil phase da a may be p o ided in his ile. These should be gi en
in weigh pe cen o he mine al soil (as sand/sil /clay pe cen ages a e) o each laye . The unc-
ion se upsoilphases.m will look o ields co esponding o he phases in ds.phase able in ds.cell;
hese phases may op ionally ha e e.g. _na i e appended. Al e na i ely, hei ea men as kine ic
o equilib ium phases may be indica ed by appending _kine ic o _eq o he phase name, e.g.,
Ens a i e_kine ic, Calci e_eq. Kine ic ea men o phases also depends on he p esence o bo h
he modynamic and kine ic pa ame e s in he phases ile; equilib ium phases equi e only he mo-
dynamic da a.
Table 2.7: Soil laye inpu pa ame e s
opdep h Top dep h cm Requi ed
bo omdep h Bo om dep h cm Requi ed
bulkden Bulk densi y g cm−3Requi ed
pa ame e de ini ion uni s
op ions
de aul pu pose
Con inued on nex page
12
Table 2.7: Soil laye inpu pa ame e s (Con inued)
sandpc Sand wg % Requi ed o pedo ans e unc ion and
na i e mine als
sil pc Sil wg % Requi ed o pedo ans e unc ion
claypc Clay wg % Requi ed o pedo ans e unc ion
TOC To al o ganic ca bon wg % Requi ed So p ion o o ganic ma e
TON To al o ganic ni ogen wg % Requi ed Decomposi ion
Mois u e pa ame e s used in e nally ha can be gi en explici ly
(de aul : use pedo ans e unc ion)
po Po osi y olume
ac ion
pedo ans . Maximum wa e con en
c Field capaci y olume
ac ion
pedo ans . Maximum wa e con en wi h-
ou low
wp Wil ing poin olume
ac ion
pedo ans . Minimum wa e con en (hy-
g oscopic wa e )
H2O Ini ial wa e olume
ac ion
ield capaci y Ini ial wa e con en a s a o
model un
Ca ion exchange
CEC Ca ion exchange
capaci y
cmolc
kg−1
soil
Toge he wi h he bulk densi y
and laye soil olume, CEC
se s he size o he (clay) ex-
change in he PHREEQC
EXCHANGE block.
Ca_exch Exchangeable Ca cmolc
kg−1
soil o
%CEC
i CEC
gi en
P o ides ini ial amoun o
an exchangeable elemen on
he exchange (clays as ep-
esen ed in he PHREEQC
EXCHANGE block). Gen-
e ally, <X>_exch o any
exchangeable elemen <X>.
Ini ial soil solu ions (de aul : lood soil wi h ainwa e )
pH_soln pH Ini ial soil solu ion pH
Ca_soln Ca mol L−1Ini ial soil solu ion elemen-
al concen a ion. Gene ally,
<X>_soln o elemen <X>,
which may also include Amm
(ammonium). P e e ably all
majo ions should be gi en.
pa ame e de ini ion uni s
op ions
de aul pu pose
Con inued on nex page
13
Table 2.7: Soil laye inpu pa ame e s (Con inued)
Solid kine ically-dissol ing na i e soil phases (RSA es . om sand/sil /clay) no eeds ocks
K eldspa _kine ic Soil mine al wg % Ini ial amoun o a na i e soil
phase ep esen ed kine ially.
Gene ally, <PHASE>_kine ic
whe e <PHASE> ma ches he
name o a phase in he phase
inpu ile (Table 2.4).
Seconda y phases o include as EQUILIBRIUM_PHASES
Calci e_eq Soil mine al mol g−1
soil Ini ial moles o equilib ium
phase <PHASE> co espond-
ing o a phase in he phase
inpu ile (Table 2.4).
pa ame e de ini ion uni s
op ions
de aul pu pose
2.5 Feeds ock PSD ile
An op ional ile con aining measu ed pa icle size dis ibu ion da a can be p o ided. This should
be a .cs ile wi h ow names in column 1. These ow headings should ma ch he s ing desc ibing
each line in he desc ip ion below. Subsequen columns con ain con en s indica ed by column 1. NB
s ings a e NOT quo ed and mus NOT con ain commas.
ield The e mus be a column heading called adius o diame e p o iding he uppe bin edge
adius o diame e . Subsequen column headings gi e he name o each eeds ock, o which
he e mus be a leas one.
label This ow con ains a ex label desc ibing each column.
uni s This ow con ains he uni s o he da a ound in each column. Uni s should be mic ons o
adius o diame e columns and weigh pe cen in each bin o eeds ock columns.
BETN2_m2g This me ada a ow con ains he BET (B unaue e al., 1938) su ace a ea o each
eeds ock, in m2 ock g−1 ock. En e 0 o he adius and diame e columns, o whe e he da a
a e missing.
geo_m2g This me ada a ow con ains he geome ic su ace a ea o each eeds ock, in m2 ock g−1
ock. En e 0 o he adius and diame e columns, o whe e he da a a e missing.
lambda This me ada a ow con ains BET/geo su ace a ea a ios. En e 0 o he adius and
diame e columns, o whe e he da a a e missing.
o ma This ow con ains he s ing % in e e y column, indica ing ha MATLAB should ea
he da a as loa ing-poin numbe s.
da a These ows con ain he measu ed da a as desc ibed by he label and uni s ows. Fo eeds ock
columns, ows con ain he weigh pe cen age o eeds ock in each bin.
14
The o ma o his ile is a legacy om ea lie model e sions which has la gely been dep eca ed
elsewhe e as i does no ake ad an age o MATLAB’s buil -in ou ines o ead .cs da a. I was
designed o make he uni s e c clea and allow inclusion o me ada a (such as he su ace a ea da a
abo e). An example o such a ile is examples/EF_da a/Lewis_ eeds ocks_PSD.cs
2.6 Time able o e en s
This ile schedules eeds ock and e ilise applica ions, i any, as well as p o iding obse ed peak
ha es da a. I mus include su icien imes amp in o ma ion o assign expe imen al day numbe
and decimal yea o each e en (see Table 2.2). The esul ing s uc u e should ha e he ields shown
in Table 2.8, some o which will be popula ed by he p ep ocesso (e.g., unindex) as no ed, bu
o he s should be columns in he ime able ile wi h headings exac ly ma ching he a iable names
shown.
The peak o , peak oo and ha es da a a e he biomasses o he c op a “peak” (when biomass
is la ges ) and ha es ime. I hese da a a e p o ided, and he c op is included in he ile ph ee-
qcdbinclude iles/SWAT_AppendixA_c op_pa ams4.cs , he p ep ocesso can gene a e biomass and
lea a ea index (LAI) cu es. This ile con ains all he da a in Appendix A o he SWAT model
(Nei sch e al., 2011) and some addi ional columns such as an al e na i e c op name and a lag
indica ing whe he he c op is a legume o no . The daily change in biomass and LAI cu es allow
he p ep ocesso o calcula e e apo anspi a ion and nu ien up ake om he soil laye s.
A limi ed se o common e ilise s will be ecognised and o mulas and mola masses will be gene -
a ed o hem, allowing hei inclusion in he PHREEQC KINETICS block. They a e assumed o be
pelle ised, bu a RATES ile will need o be w i en o hem. The e is a pa ame e “w i e RATES-
iles” which, i se (ds.pa m.w i e RATES iles=1), will allow he c ea ion o a RATES ile o hese
ea men s. Each e ilise will ha e a PHREEQC RATE ha is like he one in ph eeqcdbinclude-
iles/Pelle s.da ; hese will all be w i en o Pelle s.RATES. Run he p ep ocesso once o w i e his
ile (see Sec ion 4.2 o examples). This keywo d should be swi ched o o subsequen model uns
(ds.pa m.w i e RATES iles=0) o p e en simul aneous ba ch jobs om ying o ew i e his ile.
This keywo d appea s in he examples/EF_CLM5/de aul se ings.m ile (Sec ion 3.1).
Tillage and sowing dep hs need o be posi i e numbe s; hey should be ze o o o he ime able
en ies. The amoun o eeds ock and e ilise s in each model laye will be de e mined by he
ac ion o he laye which is abo e he gi en illage dep h.
2.7 Timese ies inpu iles
These .cs iles should con ain a leas one column ully speci ying he day o which he da a a e
alid (see Table 2.2). I mon hly da a a e p o ided hey will be assumed o be alid o calenda day
15 and will be esampled. The sampling o he da a in such a ile mus be he same o all a iables
in i ; sepa a e imese ies iles can be p o ided o da a wi h di e en sampling. Va iables (o he han
ime a iables) ha will be ecognised by he p ep ocesso a e shown in Table 2.9.
The i s line o he iles gi es he uni s. The e may be se e al accep able op ions o he uni s
(see Table 2.9). Fo e.g. hyd ological a iables, uni s may be gi en pe yea , pe mon h, o pe
15
Table 2.8: Time able s uc u e a ays
Va iable name Value o uni s Desc ip ion
expID ID s ing o desi ed expe imen (de aul : use all en ies)
ea men Name o eeds ock, e ilise o o he ea men applied
dose o m eeds ock,
pelle s
Fo m o he applied ea men . Feeds ocks a e assumed o ha e
powde ed/g anula o m. A p esen , only pelle s a e ecognised o
e ilise s.
dosage Amoun o applied ea men (blank i none)
doseuni s /ha, /m2,
kg/ha, kg/m2,
mol/ha, mol/m2
Uni s o dosage applied ( , kg o mol, pe ha o m2)
molmass g/mol Mola mass o he ea men (1 i moles we e he gi en uni s)
o mula Chemical o mula o ea men (blank i none)
eeds ock Name o eeds ock applied (blank i none, will be c ea ed om
ds. ime able. ea men and ds. eeds ock able)
c op Name o c op plan ed, ha es ed o sampled ( allow i none)
unindex Index linking ime able e en s o a ays in ds. un (will be c ea ed
when ime able ile is ead)
molpm2 mol/m2 Calcula ed moles o ea men added pe squa e me e o land
(blank i none)
dosekg kg/m2
sowdep h cm Sowing dep h
illagedep h cm Dep h o illage
peak o g m2
land To al (d y) biomass
peakabo e g m2
land Abo eg ound (d y) biomass
peak oo g m2
land Roo (d y) biomass
ha es g m2
land Yield (d y) a ha es
day. The second line gi es he name o he a iable p o ided and hese mus be alid MATLAB
a iable names as desc ibed in he beginning o Chap e 2. Va iable names ha a e meaning ul o
he p ep ocesso and model a e shown in Table 2.9.
2.8 Dep hwise ime se ies inpu iles
The name o a dep hwise ime se ies inpu ile should p o ide he name o he a iable o which he
da a a e in ended. Examples include:
dep h imese ies_soil mp.cs Soil empe a u e da a (uni s: Celsius)
These .cs iles should con ain a leas one column ully speci ying he day o which he da a a e
alid (see Table 2.2). I mon hly da a a e p o ided hey will be assumed o be alid o calenda
16
Table 2.9: Time se ies iles o o cing da a
Line 2 heading Allowed uni s Desc ip ion
< imes amp> see Table 2.2
ai emp Celsius Ai empe a u e
p ecip mm/uni ime P ecipi a ion
ET mm/uni ime To al e apo anspi a ion om
soil
anspi a ion mm/uni ime To al anspi a ion (de aul :
calcula e om ET)
e apo a ion mm/uni ime To al e apo a ion (de aul : cal-
cula e om ET)
in il a ion mm/uni ime In il a ion (de aul : use p e-
cip)
NPP gC m−1
landuni ime−1Ne p ima y p oduc i -
i y (o use peak biomass in
ime able 2.8
soil esp gC m−1
landuni ime−1Soil espi a ion (o es ima e
using biomass o NPP)
day 15. The da a will be esampled o p o ide daily alues. The sampling o he da a in such a
ile mus be he same o all a iables in i ; sepa a e imese ies iles can be p o ided o da a wi h
di e en sampling. Va iables (o he han ime a iables) ha will be ecognised by he p ep ocesso
a e shown in Table 2.9.
The i s line o he iles gi es he uni s in he i s column, assumed o be alid o all (non-
imes amp) columns. The e may be se e al accep able op ions o he uni s (see Table 2.9). Fo
(e.g.), hyd ological a iables, uni s may be gi en pe yea , pe mon h, o pe day. Columns ha
con ain laye da a may gi e he dep h o which hey a e alid and he dep h uni s, e.g. “15cm”.
The second line gi es he laye he name o he a iable p o ided and hese mus be alid MATLAB
a iable names as desc ibed in he beginning o Chap e 2. Aside om he usual ime- ela ed a iables
(Table 2.2), hese ields should con ain he s ing “lay” ollowed by an in ege gi ing he laye numbe ,
e.g., “lay1”. Any numbe o laye s can be de ined. These do no need o co espond o ei he he
laye s de ined in he ini _laye s.cs ile o he PHREEQC cells ha will be de ined in he model
(Sec ion 2.4). Fo example, he da a can come om ano he model whe e di e en s anda d laye s
a e de ined. These da a will be esampled o p o ide alues o each PHREEQC cell in he model.
2.9 Obse a ions iles
The names o iles con aining obse a ions should include he s ing “obs_”. The obse a ions iles
used o he Ene gy Fa m example uns include:
EF_obs_soil_d ains_B0_SOTON.cs D ain solu ion chemis y (ppm) o he la ge “B0” plo
17
unde soy-maize-maize o a ion.
EF_obs_soil_d ains_insi u_SOTON.cs D ain solu ion da a including empe a u e (◦C), pH,
DOC (mg C/L), PO4 (µmol/L) o he la ge plo s unde soy-maize-maize o a ion.
EF_obs_soil_lysime e s_SOTON.cs Solu ion chemis y (ppm) o he bo h la ge “B0” plo
(25 cm deep) and he small plo s (50 cm deep) unde soy-maize-maize o a ion.
EF_obs_soil_pH_la geplo s_ omIlsaKan ola_10Feb2021.cs Soil pH o dep h anges 0–
10cm and 10–30cm, o la ge plo s unde soy-maize-maize o a ion.
EF_obs_ o al_cCaMg elease_means_s de _ac ossallblocks_TiCa .cs Ca and Mg eed-
s ock elease a es ( CO2/ha) om Bee ling e al. (2024, Da ase S06).
These a e .cs iles wi h headings on line 1 and uni s on line 2. Obse a ions occupy he emaining
lines o he ile, gene ally one pe sample. They include a da ecollec ed column wi h a collec ion
da e in a MATLAB da e ime o ma (e.g., yyyy-mm-dd, dd/mm/yyyy). The collec ion da e is
necessa y because he p ep ocesso will assign each obse a ion o a pa icula index in he model
un. Likewise, iles wi h a “soil” designa ion include dep h (cm), and he p ep ocesso will assign a
cell (laye ) index o each obse a ion. Da a in iles wi h a “ o al” designa ion a e o he en i e soil
column and do no equi e a dep h.
Fo geochemical da a, elemen s can be included. The s ings “NO3”, “NH4” and “SO4” a e also
ecognised, al hough NH4 will be eplaced wi h edox-decoupled “AmmH”. Accep able uni s o
solu es include ppm, mg/L, mol/L, mmol/L, mM, umol/L and { mu}mol/L.
I is possible o include s anda d de ia ions o display pu poses. Fo example, cCa elease and
cCa elease_s de along wi h “na g” gi ing he numbe o samples a e aged a e columns in he
ile EF_obs_ o al_cCaMg elease_means_s de _ac ossallblocks_TiCa .cs which will allow calcu-
la ion o he s anda d e o o cumula i e Ca elease om eeds ock.
2.10 Files in subdi ec o y “ph eeqcdbinclude iles”
These iles include some bespoke PHREEQC da abases wi h edox-decoupled ni ogen and some
addi ional mas e species such as Ti(OH)4, U ea, and some o ganic acids. They a e based on he
Tipping_Hu ley.da da abase which ships wi h ecen e sions o PHREEQC and includes so p ion
o o ganic ma e . The basic da abase is THAmmO g.da ; THAmmO gAl.da is a e sion which
includes o ganic ma e complexa ion wi h Al om E landsson e al. (2016). THAmmO gAlP.da is
an un es ed e sion which includes some P so p ion da a.
The e a e se e al phase da a iles a ailable in he ph eeqcdbinclude iles/ di ec o y (Table 2.10).
Mos o he kine ic da a in hese iles comes om Paland i and Kha aka (2004) while mos o
he he modynamic da a a e om he THERMODDEM (Blanc e al., 2012) Mine al mass and
densi y da a a e om webmine al.com o we e calcula ed using he S oichTools package (Table 1.1).
The o igins o he da a o each phase a e indica ed in columns headed “gno es”, “ he mno es”,
“ a eno es”.
18
Table 2.10: The ph eeqcdbinclude iles di ec o y
Filename de aul ? Desc ip ion
PHREEQC da abases and da a-blocks
THAmmO g.da de aul See ex . Essen ially he PHREEQC da abase Tip-
ping_Hu ley.da wi h edox-decoupled ni ogen and a
ew addi ional species.
THAmmO gAl.da As THAmmO g.da , bu including Al complexa ion
wi h o ganic ma e (E landsson e al., 2016).
THAmmO gAlP.da UNTESTED As THAmmO gAl.da , bu including addi ional P com-
plexa ion.
SWATRATESPHASES unable de aul Ni i ica ion, deni i ica ion, o ganic ma e decompo-
si ion ollowing he SWAT2009 code (Nei sch e al.,
2011).
plan RATES5 de aul Plan nu ien up ake o indi idual elemen s. This ile
assumes plan s p e e ni a e o ammonium.
Pelle s.da de aul Pelle ised e ilise a e law (Ri ge and Peppas, 1987;
So yane e al., 2020). A copy o his ile is eplica ed o
each e ilise in he cu en di ec o y. See Sec ion ??.
*exchcoe s Al e na i e pa ame e isa ions o he EXCHANGE
block.
TRACERS De ines an in e ace and Ti species ( hese a e now
included in he da abases THAmmO g.da ).
Fakh aei_D iscoll_2015_Table1_o ganalog Al e na i e o ganic acids om Fakh aei and D iscoll
(2015).
Files use ul o de ining ds.phase able
eeds ock_phases_Lewis_THERMODDEM.cs Phases o six eeds ocks analysed by Lewis e al.
(2021).
eeds ock_phases_Lewis_SUPCRT.cs As eeds ock_phases_Lewis_THERMODDEM.cs , bu
some he modynamic da a om SUPCRT (e.g., God-
dé is e al., 2006, hei Table 2).
eeds ock_phases_pu e.cs Pu e wollas oni e and o s e i e.
seconda y_phases.cs A selec ion o seconda y phases (calci e, silica, gibbsi e,
nesquehoni e, ana ase, kaolini e, smec i e).
O he da a iles
c ops oichiome y.cs de aul De ines he elemen al s oichiome y o di e en c ops.
MMup akepa ams.cs de aul Michaelis-Men en up ake pa ame e s o di e en nu i-
en s (e.g., Roose e al., 2001).
Paland i_Kha aka_2004_pa ams.cs Wea he ing a e pa ame e s om Paland i and Kha aka
(2004) (no in o ma o ds.phase able).
pe iodic able.cs equi ed Pe iodic able da a, including alence, PHREEQC mas-
e species, and PHREEQC EXCHANGE species whe e
ele an ( ead by ge pe iodic able.m).
Chap e 3
Model pa ame e s and se ings
3.1 The op ional de aul se ings.m sc ip
Pa ame e se ings (e.g., Tables 3.1,3.2,3.4), di ec o ies o da a and ou pu s, PHREEQC cells and
o he en i ies may be de ined in an op ional MATLAB sc ip named “de aul se ings.m” in he cu en
di ec o y. All o he inpu iles o he p ep ocesso (Figu e 1.1, Table 2.1) a e plain- ex .cs iles
which will be ead om a speci ied di ec o y (ds.cs di ) which may be de ined in de aul se ings.m
o as a p ep ocesso a gumen .
The p ep ocesso execu es he op ional de aul se ings.m sc ip p io o p ocessing any command-line
a gumen s o eading any inpu .cs iles. Se ings om de aul se ings.m can he e o e be o e idden
by command-line a gumen s and da a in he .cs iles.
Pa ame e s and o he se ings can be en e ed in he de aul se ings.m ile. He e a e a ew examples
om he de aul se ings.m ile used o Ene gy Fa m example uns:
ds.dbdi =[ge en ("HOME") ’/soilmodel/ph eeqcdbinclude iles/’]; % Da abases
ds.db ile=’THAmmO gAl.da ’; % Includes E landsson e al’s Al so p ion s u
ds.ou di =[ge en ("SCRATCH") ’/EF_CLM5/’]; % W i e o his di ec o y
ds.MIXdi =’ opdown’; % MIX cell wa e s om he op down du ing pe cola ion
ds.SAme hod=’ adi ional’; % Simplis ic sh inking sphe e ou side PHREEQC
ds.EXCHme hod=’comp’; % Se EXCHANGE block explici ly (do no equilib a e)
ds.pa m.CECop =’CECca boxylic’; % Assign some CEC o ca boxylic o ganic si es
ds.pa m.se exacidi ysp=’Al’; % 5 Feb ua y 2025 Exchangeable acidi y species
3.1.1 Su ace a ea me hod
Kine ics o mine als and amo phous solid phases a e p ocessed in he RATES block o he PHREEQC
inpu ile (o included ile). Each phase has i s own cus om BASIC code in RATES o calcula ing he
change in moles o he phase wi hin PHREEQC i he e is a co esponding en y in he KINETICS
block o he PHREEQC inpu ile.
I is possible o include BASIC code o simplis ic changes in he eac i e su ace a ea o each
phase, bu a p esen he e is no coupling o PHREEQC wi h o he so wa e packages. Fo example,
19
20
Table 3.1: Pa ame e s con olling eeds ock applica ions and ex ending uns in he p ep ocesso (se upda as uc.m).
No e ha all eeds ocks can be excluded a un ime (ge ph eeqc un.m) using he “nodus ” keywo d.
pa ame e de aul uni s desc ip ion
sname Name o eeds ock (will eplace all eeds ocks in he
ime able)
epea ime able Decimal yea Vec o wi h wo yea s: Time able en ies be ween he
i s yea and he end o he ime able will be epea ed
un il he o al un ex ends h ough he second yea . Ex-
ample: Fo a ime able wi h a h ee-yea c op o a ion
and las en y in 2022, [2020.0 2071.0] will ex end he
ime able h ough 2071 by epea ing all ime able en ies
om 2020.0 h ough 2022.
no eeds ockyea s Calenda yea A ay o yea s whe e eeds ock will no be applied.
The a ay can be in any o de . Example: [2023:3:2070
2024:3:2070] will apply eeds ock e e y hi d yea s a ing
in 2025.
cease eeds ockyea Calenda yea The e will be no eeds ock applica ions a e his yea
( educes eeds ock doses o ze o and ese s dose o m o
null s ung o all eeds ocks in ime able en ies a e
his yea ).
Table 3.2: Pa ame e s con olling eac i e su ace a ea (RSA).
pa ame e de aul uni s desc ip ion
RSAscale 1 Scaling ac o o RSA o eeds ock phases
na i eRSAscale 1 Scaling ac o o RSA o na i e soil mine als
useBET 1 Use measu ed eeds ock BET su ace a ea ins ead o geome ic
P80 mm Pa icle diame e below which 80% o eeds ock pa icles a e ound
( o op ional Rosin-Rammle PSD)
RRsp ead Sp ead pa ame e o op ional Rosin-Rammle PSD (applies o eed-
s ocks)
BETN2_m2g m g−1BET su ace a ea o eeds ocks (can also be gi en wi h each eed-
s ock PSD
BET_m2g m g−1BET su ace a ea o eeds ocks (al e na i e name)
geo_m2g m g−1Geome ic su ace a ea o eeds ocks (calcula ed om PSD o om
BET/λi no gi en)
lambda λ, BET/geome ic su ace a ea a io o eeds ocks
Chap e 4
Ene gy Fa m example
4.1 Inpu iles and sc ip s
Da a iles o unning he Ene gy Fa m example simula ions a e in he examples/EF_da a/ sub-
di ec o y (Table 4.5). The emaining iles o he i e-yea uns a e in he examples/EF_CLM5/
subdi ec o y, while iles o he long uns h ough 2070 a e in he examples/EF_long uns/ subdi-
ec o y. Each o hese di ec o ies has i s own de aul se ings.m ile whe e he da abases and iles o
be used a e speci ied along wi h se ings such as illage dep h. Many o he se ings a e he same
o sho and long uns, bu he ile examples/EF_long uns/de aul se ings.m con ains pa ame e s
speci ically o he long uns which speci y how o epea da a om he ime able and how o en o
sa e in e media e iles.
Each un has i s own MATLAB sc ip (Tables 4.1,4.2, and 4.3) wi h co esponding shell sc ip o
ba ch submission. Fo example, Bcompsc ip .m is he sc ip o a baseline i e-yea un wi h he
Blue idge me abasal eeds ock (“B”) whe e no p ocesses a e excluded, and un Bcompsc ip .sh is he
co esponding shell sc ip which would be submi ed using he sba ch command. Two shell sc ip s
allow submission o he calib a ion uns (submi uning) and he “main” se o uns (submi all). No e
ha hese jobs mus be submi ed om he di ec o ies whe e hey a e ound (examples/EF_CLM5/
o examples/EF_long uns).
Table 4.1: MATLAB sc ip s and se ings o Ene gy Fa m calib a ion uns ound in he subdi ec o y exam-
ples/EF_CLM5/. All use he Blue idge eeds ock o he soy-maize-maize c op o a ion.
File RSAscale impscale Apa i e Sol e No e
iBimp5s1p1sc ip .m 1.1 5 implici Runge-Ku a bes
iBimp5s1p3sc ip .m 1.3 5 implici Runge-Ku a
Bimp1s1sc ip .m 1.0 1 implici Runge-Ku a
Bimp1s2sc ip .m 2.0 1 implici Runge-Ku a
Bimp5s1sc ip .m 1.0 5 implici Runge-Ku a
B kaps1sc ip .m 1.0 N/A kine ic Runge-Ku a c ashed
Bc aps1sc ip .m 1.0 N/A kine ic c ode
The MATLAB sc ip s o he i e-yea uns mus be un in he examples/EF_CLM5 di ec o y. They
we e designed o allow easy changing o some o he key se ings o he uns, employing a MATLAB
unc ion EF unse ings.m (in he same subdi ec o y) o se inpu s o he p ep ocesso .
27
28
Table 4.2: MATLAB sc ip s o he example i e-yea Ene gy Fa m model uns, ound in he subdi ec o y exam-
ples/EF_CLM5/.
File Feeds ock Op ion Desc ip ion
Bcompsc ip .m Blue idge baseline No p ocesses excluded.
Bnoni sc ip .m Blue idge No N-cycle No ni i ica ion, deni i ica ion o N- ixa ion.
Bclosedsc ip .m Blue idge Closed sys em CO2limi a ion
Bnososc ip .m Blue idge No so p ion Excludes so p ion including ca ion exchange.
Ccompsc ip .m con ol baseline No p ocesses excluded.
Cclosedsc ip .m con ol Closed sys em CO2limi a ion
Cnososc ip .m con ol No so p ion Excludes so p ion including ca ion exchange.
Fcompsc ip .m Fo s e i e baseline No p ocesses excluded.
Fclosedsc ip .m Fo s e i e Closed sys em CO2limi a ion
Fnososc ip .m Fo s e i e No so p ion Excludes so p ion including ca ion exchange.
Wcompsc ip .m Wollas oni e baseline No p ocesses excluded.
Wclosedsc ip .m Wollas oni e Closed sys em CO2limi a ion
Wnososc ip .m Wollas oni e No so p ion Excludes so p ion including ca ion exchange.
Table 4.3: MATLAB sc ip s o he example ex ended Ene gy Fa m model uns h ough 2070, ound in he subdi-
ec o y examples/EF_long uns/. All uns excep he con ol un ha e annual eeds ock applica ions om 2016–2020
inclusi e; Nis he o al numbe o eeds ock applica ions.
File Feeds ock NDesc ip ion
lB ep4sc ip .m Blue idge 4 No eeds ock applica ions a e 2020.
lB ep23sc ip .m Blue idge 23 A e 2020, applica ions e e y hi d yea .
lB ep55sc ip .m Blue idge 55 Annual applica ions h oughou he un.
lCsc ip .m Con ol 0 Con ol un (no used o igu es).
4.2 How o do an example Ene gy Fa m un
Fi s , unpack he a ball con aining he ARTEMIS code as desc ibed in Chap e 1. Make su e ha
he en i onmen a iables $HOME and $SCRATCH a e de ined and he MATLAB pa h de ini ion in
he s a up.m ile includes he main model di ec o y (Chap e 1). The equi ed MATLAB so wa e
lis ed in Table 1.1 mus also be ins alled and included in he MATLAB pa h.
On he command line, i is possible o simply call he MATLAB sc ip (e.g., Bcompsc ip .m) on he
MATLAB command line and bo h he p ep ocesso and he model should un, ying up MATLAB
o a leas hal an hou :
> Bcompsc ip
> = weak uns( ,1); % C ea es a ew ex a ields and changes d ainage cell
He e, he “>” symbol is he MATLAB p omp and should no be yped. The esul s o he p e-
p ocesso will be in s uc u e “ds”, and he esul s o he main model will be in s uc u e “ ”. The
weak uns.m pos p ocesso was designed o he Ene gy Fa m uns and is use ul o making he
igu es and ables o he ARTEMIS GMD pape . I will calcula e some o e all luxes, weak he
un labels, emo e some i ele an obse a ional da a, and ese he laye o compa ison wi h he
d ainage wa e . I s a gumen s a e he ou pu s uc u e om ge ph eeqc un.m and a lag o add he
base sa u a ion o he un label (no longe used, so his se ing makes no di e ence).
Use o EF unse ings.m makes he calling sequences o he p ep ocesso and he main model less
ob ious. To do he same un speci ying he calling sequences:
29
> % se upda as uc.m is he p ep ocesso
> ds=se upda as uc(’SAme hod’,’ adi ional’,’ sname=Blue idge’,...
’EXCHme hod’,’comp’,’pa m=igno eSws’,1,...
’pa m=implici ’,{’Ti ani e_blue idge’,’Hyd oxyapa i e_blue idge’},...
’ unname= Bcomp’,’ unlab=Ap+Ti implici ’,’pa m=RSAscale’,1.1,...
’pa m=impscale’,5);
> Bcomp=ge ph eeqc un(ds,’nyea s’,5); % main model (will ake >30min o un)
> Bcomp= weak uns( Bcomp,1); % bespoke pos p ocesso o Ene gy Fa m examples
He e, he MATLAB unc ion se upda as uc.m is he p ep ocesso , which will p in a numbe o
hings abou he un se up. I e u ns a da a s uc u e “ds” which should be passed o he main
model unc ion ge ph eeqc un.m; a copy o his da a s uc u e (possibly wi h un ime changes) will
be included in he ou pu om ge ph eeqc un.m as a subs uc u e. In he abo e example, Bcomp
is he ARTEMIS ou pu s uc u e; Bcomp.ds con ains he da a s uc u e om he p ep ocesso .
In hese uns (and indeed all uns included in he example), he soil mois u e om he CLM5
model is excluded (igno eSws) because i makes he soils e y we and uins he ni ogen-cycling
ou pu s. Also, he CLM5 mois u e can g ea ly exceed ield capaci y and a imes app oach o al
sa u a ion. ARTEMIS uns would likely be in alid gi en he limi a ions associa ed wi h gas di usion
in ARTEMIS.
The main unc ion ge ph eeqc un.m will i s p in some in o ma ion abou he un se up (such as
which p ocesses a e included and which a e excluded), and i will hen call PHREEQC o each day
o he un, ead he esul ing selec ed-ou pu ile (.sel), add hose esul s o he ARTEMIS ou pu
a ays, and p in a line s a ing whe he ha call o PHREEQC was success ul o no along wi h
a imes amp and he numbe o PHREEQC wa nings. A he end o he un, he numbe o days
p ocessed, he numbe o PHREEQC wa nings and he elapsed ime will be p in ed along wi h he
name o he ou pu .ma ile (pa h o ou pu iles edac ed):
...
Bcomp i e a ion 1825 Day 1825 s a ing: 06-Sep-2025 19:20:38: PHREEQC OK 1 wa nings
(ge o al ields/ge s elease) Ti ani e_blue idge is no a kine ic phase
so will be excluded om eeds ock elease a ays
(ge o al ields/ge s elease) Hyd oxyapa i e_blue idge is no a kine ic phase
so will be excluded om eeds ock elease a ays
Run ime o Bcomp (1825 days, 1828 wa nings): 37.2153 minu es
sa ema <X>/ Bcomp_Blue idge_Jan2016_Dec2020_06-Sep-2025.ma a 06-Sep-2025 19:20:40
In his case, i ani e and hyd oxyapa i e a e lumped oge he as an implici phase, which is included
in he eeds ock elemen al elease a ays. The 1825 lines o he PHREEQC i e a ions, along wi h
he o he p in ed s a emen s, will soon ill up he MATLAB command window. I unning in ba ch
mode, he esul ing slu m-xxx.ou ile will con ain all hose ou pu lines which can be pe used as
desi ed. This is especially use ul i he model c ashes, as o he calib a ion un B kaps1 (pa h
edac ed):
...
sa ema <X>/ B kaps1_Blue idge_Jan2016_Dec2020_06-Sep-2025.ma a 06-Sep-2025 19:22:02
<X>/ B kaps1_Blue idge_Jan2016_Dec2020_06-Sep-2025.ma and
30
<X>/ B kaps1_Blue idge_Jan2016_Dec2020_06-Sep-2025.dmp_i e 1095 SAVED a i e a ion 1095
...
B kaps1 i e a ion 1269 Day 1269 s a ing: 06-Sep-2025 21:10:20: PHREEQC OK 6 wa nings
B kaps1 i e a ion 1270 Day 1270 s a ing: 06-Sep-2025 21:15:13: ERROR:
Bad RK s eps > 1000. Please dec ease ( ime)s ep o inc ease -bad_s ep_max.
5 wa nings
E o p ocessing B kaps1 a i e a ion 1270 unday=1270
Run ime o B kaps1 (ERROR a day 1270, 537 wa nings): 153.299 minu es
sa ema <X>/ B kaps1_Blue idge_Jan2016_Dec2020_06-Sep-2025.ma a 06-Sep-2025 21:18:53
The pa ial un can be loaded, examined and displayed wi h he o he uns.
One can see when he un c ashed, when he las ou pu ile was sa ed, and he e o h own by
PHREEQC. T oubleshoo ing PHREEQC e o s can be icky bu ARTEMIS does include a ew
op ions o weaking he PHREEQC nume ical me hod. He e, he e o sugges s ha PHREEQC
eached he maximum numbe o a emp s o in eg a e he sys em o kine ic equa ions du ing a
eac ion s ep. The PHREEQC de aul o he numbe o s eps (PHREEQC se ing -bad_s ep_max)
is 500 s eps. This pa ame e can be adjus ed when calling ge ph eeqc un using he ARTEMIS
bads epmax keywo d, and he numbe o e enly-di ided s eps o KINETICS (cu en ly 3) can be
adjus ed wi h he ARTEMIS nkins eps keywo d. Howe e , he Runge-Ku a me hod is no ideal o
uns wi h kine ic apa i e as he se o equa ions o be sol ed is oo s i (Pa khu s and Appelo, 2013,
pages 110–111).
Mos pa ame e s can be se as shown o RSAscale and impscale (Table 3.2). The un label is o he
igu e legends; ou pu iles om ge ph eeqc un.m will ha e names s a ing wi h he gi en unname
(Table 4.4). These will include he s anda d PHREEQC inpu , ou pu , selec ed ou pu and dump
iles (Pa khu s and Appelo, 2013) o he i s and las days o he un, along wi h some in e media e
days speci ied by he sa e in e al (de aul : 365 days). To sa e space, hese in e media e iles ha e
no been e ained in he a balls con aining he o he ou pu iles.
The MATLAB s uc u e e u ned by ge ph eeqc un.m will be sa ed in he simila ly-named ile wi h
a .ma ex ension, which can be loaded on he MATLAB command line. The .ma ile will con ain
esul s co e ing he en i e un.
Table 4.4: Ou pu iles o he Ene gy Fa m example un “ Bcomp”.
File Desc ip ion
Bcomp_Blue idge_Jan2016_Dec2020_03-Sep-2025.dmp_ini PHREEQC dump ile o he i s day o he un.
Bcomp_Blue idge_Jan2016_Dec2020_03-Sep-2025.dmp_i e nPHREEQC dump ile o day no he un.
Bcomp_Blue idge_Jan2016_Dec2020_03-Sep-2025.dmp PHREEQC dump ile o he las day o he un.
Bcomp_Blue idge_Jan2016_Dec2020_03-Sep-2025.in_ini PHREEQC inpu ile o he i s day o he un.
Bcomp_Blue idge_Jan2016_Dec2020_03-Sep-2025.in_i e nPHREEQC inpu ile o day no he un.
Bcomp_Blue idge_Jan2016_Dec2020_03-Sep-2025.in PHREEQC inpu ile o he las day o he un.
Bcomp_Blue idge_Jan2016_Dec2020_03-Sep-2025.ou _ou i PHREEQC ou pu ile o he i s day o he un.
Bcomp_Blue idge_Jan2016_Dec2020_03-Sep-2025.ou _i e nPHREEQC ou pu ile o day no he un.
Bcomp_Blue idge_Jan2016_Dec2020_03-Sep-2025.ou PHREEQC ou pu ile o he las day o he un.
Bcomp_Blue idge_Jan2016_Dec2020_03-Sep-2025.sel_ini PHREEQC selec ed-ou pu ile o he i s day o he un.
Bcomp_Blue idge_Jan2016_Dec2020_03-Sep-2025.sel_i e nPHREEQC selec ed-ou pu ile o day no he un.
Bcomp_Blue idge_Jan2016_Dec2020_03-Sep-2025.sel PHREEQC selec ed-ou pu ile o he las day o he un.
Bcomp_Blue idge_Jan2016_Dec2020_28-Aug-2025.sc eenou PHREEQC sc een ou pu om he las day o he un.
Bcomp_Blue idge_Jan2016_Dec2020_29-Aug-2025.ma MATLAB ile con aining ARTEMIS ou pu o he en i e un.
31
O he se ings will be ound in he MATLAB sc ip s. O iginally, uns allowing PHREEQC o equi-
lib a e he EXCHANGE block (EXCHme hod=’eq’) we e done o compa ison wi h uns speci ying
he EXCHANGE composi ion explici ly (EXCHme hod=’comp’). Runs excluding he UAN e ilis-
e s and plan nu ien up ake we e also unde aken bu hese p ocesses p o ed o ha e mino e ec s
on he key ou pu s o in e es (Ca and Mg elease om eeds ock and solu ion chemis y).
Table 4.5: Da a iles o he example Ene gy Fa m model uns, ound in he subdi ec o y examples/EF_da a/.
File Desc ip ion
EF_CLM5_2016-2070_ imese ies_CO2a m.cs A mosphe ic CO2 om CLM5 o long uns.
EF_CLM5_2016-2070_ imese ies_in il_ET.cs Hyd ological o cings om CLM5 o long uns.
EF_CLM5_2016-2070_ imese ies_soil esp.cs Soil espi a ion om CLM5 o long uns.
EF_CLM5_2016-2070_ imese ies_soil emp.cs Soil empe a u e om CLM5 o long uns.
EF_CLM5_ imese ies_CO2a m.cs A mosphe ic CO2 om CLM5 o i e-yea uns.
EF_CLM5_ imese ies_hyd o.cs Hyd ological o cings om CLM5 o i e-yea
uns.
EF_CLM5_ imese ies_in il_ET.cs In il a ion om CLM5 o i e-yea uns.
EF_CLM5_ imese ies_soil esp.cs Soil espi a ion om CLM5 o i e-yea uns.
EF_CLM5_ imese ies_soil emp.cs Soil empe a u e om CLM5 o i e-yea uns.
EF_maize_B0_ ime able.cs Time able o e en s a he Ene gy Fa m.
EF_obs_soil_d ains_B0_SOTON.cs D ainage majo ion chemis y o he la ge “B0”
me abasal - ea ed plo o he Ene gy Fa m
soy-maize-maize o a ion.
EF_obs_soil_d ains_insi u_SOTON.cs D ainage pH o he la ge “B0” me abasal -
ea ed plo o he Ene gy Fa m soy-maize-
maize o a ion.
EF_obs_soil_lysime e s_SOTON.cs Majo ion chemis y measu ed in lysime e s o
he la ge “B0” me abasal - ea ed plo (25 cm
dep h) and he ou small plo s (50 cm dep h)
o he Ene gy Fa m soy-maize-maize o a ion.
EF_obs_soil_pH_la geplo s_ omIlsaKan ola_pe scomm_10Feb2021.cs Soil pH o he la ge “B0” me abasal - ea ed
plo o he Ene gy Fa m soy-maize-maize o a-
ion.
EF_obs_ o al_cCaMg elease_means_s de _ac ossallblocks_TiCa .cs Ca and Mg wea he ing elease da a om Bee -
ling e al. (2024).
ini laye s_Flanagan_EF_0-183cm.cs Soil laye da a o one o he mos common soils
in he Ene gy Fa m soy-maize-maize o a ion
plo s. This ile con ains me ada a including he
da a sou ces.
Lewis_ eeds ocks_PSD.cs Pa icle size dis ibu ion da a o he six eed-
s ocks analysed by Lewis e al. (2021), including
he Blue idge me abasal applied a he Ene gy
Fa m.
ainwa e _Champaign_NTN_2021_ini ialChemis y.cs Rainwa e chemis y ep esen an a e age o
he ainwa e collec ed o e he pe iod s a ing
2020-11-17 15:15 and ending 2020-11-24 15:15 by
he Na ional T ends Ne wo k unde he Na ional
A mosphe ic Deposi ion P og am o Cham-
paign, IL, USA, downloaded on F iday 25 June
2021. I all gi en species a e included, pe cen
e o in he cha ge balance o his solu ion is
0.0351%.
Richland_loess_Ba es own_ ill_phases_combined.cs Da a o mine als associa ed wi h he pa en
ma e ials a he Ene gy Fa m.
32
4.3 Loading and displaying he Ene gy Fa m esul s
The i s hing o do is o unpack he a balls con aining he model uns. These sc ip s assume ha
ARTEMIS_GMD_EF_CLM5_ou pu iles_Sep-2025. a .gz was unpacked in examples/EF_CLM5/
and ARTEMIS_GMD_EF_long uns_ou pu iles_Sep-2025. a .gz in examples/EF_long uns/, c e-
a ing ou pu iles subdi ec o ies.
The e a e wo MATLAB sc ip s which I used o load he ou pu s o mos o he i e-yea uns:
EF_CLM5/load uning uns.m Loads he i e-yea -long uns done o calib a ion wi h he blue idge
eeds ock, c ea ing MATLAB cell a ay o s uc u es “ ”.
EF_CLM5/loadmain uns.m Loads he i e-yea -long uns done wi h di e en eeds ocks, c ea -
ing MATLAB cell a ay o s uc u es “ ”
The e a e i e MATLAB sc ip s which we e used o c ea e he igu es o he GMD pape , wo o
which also load ARTEMIS uns:
EF_CLM5/Flanaganlaye diag am.m C ea es GMD loca ion map and laye diag am (load he
main uns wi h EF_CLM5/loadmain uns.m i s as he laye da a will be ead om 1,1)
EF_CLM5/ uning igs.m GMD igu es ela ed o calib a ion wi h he blue idge eeds ock and
loaded wi h EF_CLM5/load uning uns.m
EF_CLM5/ eeds ock igs.m GMD igu es ela ed o uns done wi h di e en eeds ocks and
loaded wi h EF_CLM5/loadmain uns.m
EF_CLM5/BRnoni igs.m Loads uns and c ea es GMD igu es ela ed o he e ec o he ni-
ogen cycle
EF_CLM5/long e m igs.m Loads uns co e ing he pe iod 2016–2070 and c ea es GMD igu es
and ables ela ed o lag imes and e ec s o con inued ea men s wi h he blue idge eeds ock
Each o hese sc ip s con ains MATLAB swi ches which de e mine which igu es will be made. These
should be edi ed o choose which igu es o look a . They also con ain a a iable “png” which
de e mines whe he he igu es a e sa ed o no ; his is se o ze o (do no sa e) i i does no exis .
4.4 Ene gy Fa m calib a ion uns
ARTEMIS was calib a ed such ha he Ca and Mg elease om eeds ock was wi hin one s anda d
e o o elease a es de i ed by Bee ling e al. (2024) using he “TiCAT” me hod (Ree shemius e al.,
2023) whe eby changes in ca ion con en o soil samples a e compa ed o an “immobile” elemen , in
his case Ti. Li le o no a emp was made o calib a e o he ARTEMIS ou pu s, o he han o
change a soil mois u e h eshold o deni i ica ion.
As a i s a emp , all eeds ock phases wi h a e laws we e modelled kine ically using a e laws
om Paland i and Kha aka (2004), and one Ca-bea ing phase wi h no kine ic o he modynamic
pa ame e s ( i ani e) was igno ed. PHREEQC has wo nume ical sol e s (Pa khu s and Appelo,
33
2013). Using he de aul Runge-Ku a (RK) sol e wi h six ime subin e als and h ee kine ic s eps,
PHREEQC c ashed wi h con e gence e o s be o e comple ing he i e-yea un. Ano he un wi h
PHREEQC’s implici s i o dina y di e en ial equa ion sol e “c ode” did no c ash, bu ook 258
minu es o un o 1825 days.
Cumula i e Mg elease was accep able in he kine ic c ode un, bu Ca elease was low compa ed
o TiCAT (Fig. 4.1), sugges ing a p oblem wi h he ep esen a ion o one o mo e Ca-bea ing eed-
s ock phases. One such phase is i ani e, which has no kine ic a e pa ame e s. Ano he is apa i e
(Table 4.6), which con ains phospho us. In bo h kine ic uns, apa i e wea he ed oo quickly such
ha P was eleased and leached du ing he allow seasons a e eeds ock applica ion, lea ing li le
o none o plan up ake du ing he ollowing g owing seasons (Sec ion 4.4.1). ARTEMIS includes
kine ic plan P up ake and P elease om decomposing o ganic ma e , and he PHREEQC da abase
includes phospha e specia ion and so p ion o hyd ous e ic oxides. Howe e , P can also so b o a
a ie y o o he su aces (e.g., Jalali e al., 2022, and e e ences in hei Table S1) which could po-
en ially educe he apa i e sa u a ion index and inc ease i s wea he ing a es, bu would also e ain
mo e P in he sys em. P cycling in ARTEMIS could be imp o ed wi h su aces speci ically de eloped
and calib a ed o he Ene gy Fa m soils, bu his is ou side he scope o he p esen example.
Table 4.6: Mine alogy o he “blue idge” me abasal (Lewis e al., 2021) applied a he Ene gy Fa m (2016–2019).
All pa ame e s o hese phases a e gi en in he inpu ile “ eeds ock_phases_Lewis_THERMODDEM.cs ”, in-
cluded wi h he so wa e and example iles. The modynamic da a a e om he THERMODDEM da abase (BRGM,
2020; Blanc e al., 2012). Qua z kine ic pa ame e s a e om Bands a and B an ley (2008) and Tes e e al.
(1994); o he phases ollow Paland i and Kha aka (2004). The “kine ics” column shows how each phase was ea ed
in he inal calib a ed model un.
Phase Fo mula Weigh % kine ics
Albi e NaAlSi3O819.6 i e e sible
Fe oac inoli e Ca2(Mg0.75Fe0.25)5Si8O22(OH)211.6 i e e sible
Epido e Ca2Al2(Al0.84Fe0.16)(SiO4)(Si2O7)O(OH) 25.6 i e e sible
Chlo i e (Mg0.63Fe0.37)5Al(Si3Al)O10(OH)836.3 i e e sible
Qua z SiO25.2 e e sible
Ti ani e CaTiSiO51.7 implici
Musco i e KAl3Si3O10(OH)1.8F0.23.3345 i e e sible
Apa i e Ca5(PO4)3(OH)11.9325 implici
As expec ed, a es un igno ing bo h apa i e and i ani e p oduced poo esul s, wi h cumula i e
Ca+Mg elease o e i e CO2ha−1 oo low (Fig. 4.1). Al e na i ely, apa i e can be ea ed as
an implici phase along wi h i ani e, so ha hey e ec i ely dissol e in andem wi h he kine ic
eeds ock phases. Using his implici “phase” wi hou any scaling, howe e , does no imp o e he
RMSEs, and inc easing he whole- eeds ock ini ial eac i e su ace a ea using he RSAscale keywo d
(Table 3.2) imp o es model Ca elease a he expense o Mg elease. Ins ead, a combina ion o
sligh ly inc eased eeds ock RSA scaling by a ac o o 1.1 and subs an ially inc eased “implici ”
phase dissolu ion a e by a ac o o i e p oduced esul s deemed accep able o he pu poses o his
example s udy (Fig. 4.1, blue cu e). Runs whe e apa i e was no a kine ic phase ook app oxima ely
40 minu es o un on he high-pe o mance compu e a he Uni e si y o She ield.
34
Figu e 4.1: Compa ison be ween model calib a ion uns (solid lines) and TiCAT (blue do s, Bee ling e al., 2024;
Ree shemius e al., 2023) Ca+Mg (a), Ca (b) and Mg (c) elease a es om he “blue idge” eeds ock a he Ene gy
Fa m (2016–2019). The hick blue solid line ep esen s he bes model un used in subsequen expe imen s in he
cu en s udy.
4.4.1 Phospho us dynamics du ing calib a ion uns
The calib a ing exe cise se ed o unde sco e key limi a ions wi h P cycling in ARTEMIS, which can
be obse ed in he opsoil model P pools (Fig. 4.2). Whe e ea ed as a kine ic phase, as -wea he ing
apa i e eleased P du ing he allow season. P was hen leached be o e i could be aken up by plan s,
sugges ing insu icien P so p ion. Feeds ock apa i e was he main sou ce o P in hese uns, wi h only
small amoun s being eleased om o ganic ma e . Unsu p isingly, i apa i e was igno ed, he e was
oo li le P in he sys em. Wi h “implici ” apa i e and i ani e, plan P up ake imp o ed, bu P was
eleased e y slowly and apa i e accumula ed in he soil. This e ec is pa ially mi iga ed i apa i e
(and i ani e) dissolu ion was inc eased ela i e o dissolu ion o kine ically-wea he ing phases. The
bes esul s we e ob ained by inc easing he implici dissolu ion by a ac o o i e, oge he wi h a
modes 1.1- old inc ease in he ini ial eac i e su ace a ea o he whole eeds ock (Fig. 4.2a). Plan
P demand was howe e no qui e sa is ied du ing hese i e-yea uns (Fig. 4.3), which also sugges s
insu icien P e en ion in he model soil.
35
Figu e 4.2: Topsoil P pools du ing calib a ing uns. He e, g ey is he P emaining in eeds ock, blue is P in solu-
ion, and g een is P aken up by he plan . Apa i e dissol es quickly enough o acili a e plan P up ake in he un
wi h he bes RMSE o Ca+Mg elease om eeds ock (a). O he uns wi h slowe “implici ” apa i e dissolu ion
p o ide less P o plan s and allow apa i e o accumula e (b, c). Wi hou apa i e (c), o wi h apa i e ea ed as a
as -wea he ing kine ic phase (d,e), he e is li le P a ailable o plan up ake.
36
Figu e 4.3: To al cumula i e P up ake om all soil laye s by plan s du ing calib a ing uns. He e, he hin blue line
is o al plan P demand, and he hick lines a e P up ake cu es o he calib a ing uns.
43
Figu e 5.3: The main unc ion ge ph eeqc un.m w i es he i s PHREEQC inpu ile which ollows he sequence o
e en s shown.
shows acked ields.m C ea es a s ack o plo s, calling show un, showCEC, showsu aces, o show-
cellel balance o mul iple cells o mul iple uns.
showAlk Shows he majo ions con ibu ing o he alkalini y balance.
44
Figu e 5.4: The main unc ion ge ph eeqc un.m w i es he PHREEQC inpu ile which ollows he sequence o
e en s shown.
Bibliog aphy
Bands a, Joel Z and Susan L B an ley (2008). Da a i ing echniques wi h applica ions o mine al dissolu ion kine ics.
In: Kine ics o wa e - ock in e ac ion. Sp inge , 211–257 (ci . on p. 33)
Bee ling, Da id J, Dimi a Z Epiho , Ilsa B Kan ola, Michael D Mas e s, Tom Ree shemius, Noah J Plana sky,
Ch is ophe T Reinha d, Jacob S Jo dan, Sa ah J Tho ne, James Webe , e al. (2024). Enhanced wea he ing in he
US Co n Bel deli e s ca bon emo al wi h ag onomic bene i s. P oceedings o he Na ional Academy o Sciences
121, e2319436121. doi:h ps://doi.o g/10.1073/pnas.2319436121 (ci . on pp. 2,17,31,32,34)
Bee ling, Da id J, Eu ipides Kan zas, Ma k R Lomas, Pe e Wade, Ra ael M Eu asio, Phil Ren o h, Binoy Sa ka ,
M G ace And ews, Rachael H James, Ch is ophe R Pea ce, e al. (2020). Po en ial o la ge-scale CO2 emo al ia
enhanced ock wea he ing wi h c oplands. Na u e 583, 242–248. doi:h ps://doi.o g/10.1038/s41586-020-
2448-9 (ci . on pp. 3,4,21,23)
Bee ling, Da id J, Eu ipides P Kan zas, Ma k R Lomas, Lyla L Taylo , Shuang Zhang, Yoshiki Kanzaki, Ra ael M
Eu asio, Phil Ren o h, Jean-F ancois Mecu e, Hec o Polli , e al. (2025). T ans o ming US ag icul u e o ca bon
emo al wi h enhanced wea he ing. Na u e, 1–10. doi:10.1038/s41586-024-08429-2 (ci . on p. 4)
Ben-Noah, Ilan and Shmulik P F iedman (2018). Re iew and e alua ion o oo espi a ion and o na u al and ag icul-
u al p ocesses o soil ae a ion. Vadose Zone Jou nal 17, 1–47. doi:h ps://doi.o g/10.2136/ zj2017.06.0119
(ci . on p. 22)
Blanc, Ph, A naul Lassin, Pa ice Pian one, Mohamed Aza oual, Nicolas Jacqueme , A Fabb i, and E ic C Gauche
(2012). The moddem: A geochemical da abase ocused on low empe a u e wa e / ock in e ac ions and was e ma-
e ials. Applied geochemis y 27, 2107–2116. doi:10.1016/j.apgeochem.2012.06.002 (ci . on pp. 17,33)
BRGM (2020). The moddem PHREEQC da abase. Ve sion 1.10, 15 Decembe 2020 (ci . on p. 33)
B unaue , S ephen, Paul Hugh Emme , and Edwa d Telle (1938). Adso p ion o gases in mul imolecula laye s.
Jou nal o he Ame ican chemical socie y 60, 309–319 (ci . on p. 13)
Ce ling, Thu e E (1991). Ca bon dioxide in he a mosphe e: e idence om Cenozoic and Mesozoic paleosols. Ame ican
Jou nal o Science;(Uni ed S a es) 291, 377–400. doi:h ps://doi.o g/10.2475/ajs.291.4.377 (ci . on pp. 22,
25,38)
Choudhu y, Bhaska J (2001). Modeling adia ion-and ca bon-use e iciencies o maize, so ghum, and ice. Ag icul u al
and Fo es Me eo ology 106, 317–330. doi:h ps://doi.o g/10.1016/S0168-1923(00)00217-3 (ci . on p. 25)
Cosby, BJ, GM Ho nbe ge , RB Clapp, and ToR Ginn (1984). A s a is ical explo a ion o he ela ionships o soil
mois u e cha ac e is ics o he physical p ope ies o soils. Wa e Resou ces Resea ch 20, 682–690. doi:h ps:
//doi.o g/10.1029/WR020i006p00682 (ci . on p. 11)
Du ene, E, R Ochs, and B Saugie (1990). Pho osyn hèse e p oduc ion du palmie à huile en ela ion a ec les ac eu s
clima iques. Oléagineux 45, 8–9 (ci . on p. 25)
Dzombak, Da id A and F ancois MM Mo el (1990). Su ace complexa ion modeling: hyd ous e ic oxide. John Wiley
& Sons (ci . on p. 21)
E landsson, M, EH Oelke s, Ke in Bishop, H S e d up, S Belyazid, JLJ Ledesma, and SJ Köhle (2016). Spa ial and
empo al a ia ions o base ca ion elease om chemical wea he ing on a hillslope scale. Chemical Geology 441,
1–13. doi:h ps://doi.o g/10.1016/j.chemgeo.2016.08.008 (ci . on pp. 17,18)
45
46
Fakh aei, Habibollah and Cha les T D iscoll (2015). P o on and aluminum binding p ope ies o o ganic acids in
su ace wa e s o he no heas e n US. En i onmen al Science & Technology 49, 2939–2947. doi:h ps://doi.
o g/10.1021/es504024u (ci . on p. 18)
Goddé is, Y es, Louis M F ançois, Anne P obs , Jacques Scho , Da id Moncoulon, Da id Laba , and Daniel Vi ille
(2006). Modelling wea he ing p ocesses a he ca chmen scale: The WITCH nume ical model. Geochimica e Cos-
mochimica Ac a 70, 1128–1147. doi:10.1016/j.gca.2005.11.018 (ci . on p. 18)
Ho mann, Muni P, A Cas aneda Ve a, MT Van Wijk, Ken E Gille , Thomas Obe hue , Ch is ophe Donough, and
An hony M Whi b ead (2014). Simula ing po en ial g ow h and yield o oil palm (Elaeis guineensis) wi h PALMSIM:
Model desc ip ion, e alua ion and applica ion. Ag icul u al Sys ems 131, 1–10. doi:h ps://doi.o g/10.1016/
j.agsy.2014.07.006 (ci . on p. 25)
Jalali, Mohsen, Elham Ami abadi Fa ahani, and Mahdi Jalali (2022). Simula ing phospho us leaching om wo ag i-
cul u al soils as a ec ed by di e en a es o phospho us applica ion based on he geochemical model PHREEQC.
En i onmen al Moni o ing and Assessmen 194, 164. doi:10.1007/s10661-022-09828-6 (ci . on p. 33)
Kan ola, Ilsa B, Elena Blanc-Be es, Michael D Mas e s, Ellio Chang, Alison Ma klein, Cai lin E Moo e, Adam on
Haden, Ca l J Be nacchi, Adam Wol , Dimi a Z Epiho , e al. (2023). Imp o ed ne ca bon budge s in he US
Midwes h ough di ec measu ed impac s o enhanced wea he ing. Global Change Biology 29, 7012–7028. doi:
h ps://doi.o g/10.1111/gcb.16903 (ci . on p. 2)
Kan ola, Ilsa B, Michael D Mas e s, Da id J Bee ling, S ephen P Long, and E an H DeLucia (2017). Po en ial o
global c oplands and bioene gy c ops o clima e change mi iga ion h ough deploymen o enhanced wea he ing.
Biology le e s 13, 20160714. doi:h ps://doi.o g/10.1098/ sbl.2016.0714 (ci . on p. 2)
Kan zas, E, M Val Ma in, M Lomas, R Eu asio, P Ren o h, A Lewis, L Taylo , J-F Mecu e, H Polli , P Ve coulen,
N Vakili a d, P Holden, N Edwa ds, L Koh, N Pidgeon, S Banwa , and D Bee ling (2022). Subs an ial ca bon
d awdown po en ial om enhanced ock wea he ing in he Uni ed Kingdom. Na u e Geoscience 15, 382–389. doi:
h ps://doi.o g/10.1038/s41561-022-00925-2 (ci . on pp. 4,21,23)
Kelland, Michael, Pe e Wade, Amy Lewis, Lyla Taylo , Binoy Sa ka , G ace And ews, Ma k Lomas, Anne Co on,
Simon Kemp, Rachael James, Ch is Pea ce, Sue Ha ley, Ma k Hodson, Jona han R Leake, S e e Banwa , and
Da id Bee ling (2020). Inc eased yield and CO2seques a ion po en ial wi h he C4ce eal So ghum bicolo cul i a ed
in basal ic ock dus -amended ag icul u al soil. Global Change Biology 26, 3658–3676. doi:10.1111/gcb.15089
(ci . on pp. 2,7)
Ki schbaum, Miko UF and Rowena Muelle (2001). Ne Ecosys em Exchange: Wo kshop P oceedings. Coope a i e
Resea ch Cen e o G eenhouse Accoun ing (ci . on p. 25)
Lamade, Emmanuelle, Na cisse Djégui, and Philippe Le e me (1996). Es ima ion o ca bon alloca ion o he oo s
om soil espi a ion measu emen s o oil palm. Plan and soil 181, 329–339. doi:h ps://doi.o g/10.1007/
BF00012067 (ci . on p. 25)
Law ence, D, R Fishe , C Ko en, K Oleson, S Swenson, and M Ve ens ein (23 Ma ch 2020). Technical desc ip ion
o e sion 5.0 o he Communi y Land Model (CLM). Na ional Cen e o A mosphe ic Resea ch. PO Box 3000,
Boulde , CO 80307-300 (ci . on pp. 2,41)
Lewis, Amy L, Binoy Sa ka , Pe e Wade, Simon J Kemp, Ma k E Hodson, Lyla L Taylo , Kok Loong Yeong, Kalu
Da ies, Paul N Nelson, Michael I Bi d, e al. (2021). E ec s o mine alogy, chemis y and physical p ope ies o
basal s on ca bon cap u e po en ial and plan -nu ien elemen elease ia enhanced wea he ing. Applied Geochem-
is y 132, 105023. doi:10.1016/j.apgeochem.2021.105023 (ci . on pp. 18,31,33)
Nei sch, Susan L, Je ey G A nold, Jim R Kini y, and Jimmy R Williams (2011). Soil and Wa e Assessmen Tool
heo e ical documen a ion e sion 2009. Tech. ep. Texas Wa e Resou ces Ins i u e (ci . on pp. 2,14,18,22)
Noilhan, J and P Laca è e (1995). GCM g id-scale e apo a ion om mesoscale modeling. Jou nal o Clima e 8, 206–
223. doi:h ps://doi.o g/10.1175/1520-0442(1995)008<0206:GGSEFM>2.0.CO;2 (ci . on p. 11)
Paland i, James L and Yousi K Kha aka (2004). A compila ion o a e pa ame e s o wa e -mine al in e ac ion
kine ics o applica ion o geochemical modeling. Tech. ep. US Geological Su ey. doi:h ps://doi.o g/10.
3133/o 20041068 (ci . on pp. 17,18,32,33)
Pa khu s , D.L. and C.A.J. Appelo (2013). Desc ip ion o inpu and examples o PHREEQC e sion 3—A compu e
p og am o specia ion, ba ch- eac ion, one-dimensional anspo , and in e se geochemical calcula ions. Techniques
and Me hods. U.S. Geological Su ey. Chap. A43, 497 (ci . on pp. 1,2,24,30,32)
47
Raich, James W and William H Schlesinge (1992). The global ca bon dioxide lux in soil espi a ion and i s ela ionship
o ege a ion and clima e. Tellus B 44, 81–99. doi:h ps://doi.o g/10.1034/j.1600- 0889.1992. 01-1-
00001.x (ci . on p. 25)
Ree shemius, Tom, Mike E Kelland, Jacob S Jo dan, Isabelle R Da is, Rocco D’Ascanio, Bo iana Kalde on-Asael, Dan
Asael, T Jespe Suh ho , Dimi a Z Epiho , Da id J Bee ling, e al. (2023). Ini ial alida ion o a soil-based mass-
balance app oach o empi ical moni o ing o enhanced ock wea he ing a es. En i onmen al Science & Technology
57, 19497–19507. doi:10.1021/acs.es .3c03609 (ci . on pp. 32,34)
Ren o h, Phil and Gideon Hende son (2017). Assessing ocean alkalini y o ca bon seques a ion. Re iews o Geo-
physics 55, 636–674. doi:h ps://doi.o g/10.1002/2016RG000533 (ci . on p. 7)
Ri ge , Philip L and Nikolaos A Peppas (1987). A simple equa ion o desc ip ion o solu e elease I. Fickian and
non- ickian elease om non-swellable de ices in he o m o slabs, sphe es, cylinde s o discs. Jou nal o Con olled
Release 5, 23–36. doi:10.1016/0168-3659(87)90034-4 (ci . on p. 18)
Roose, Tiina, AC Fowle , and PR Da ah (2001). A ma hema ical model o plan nu ien up ake. Jou nal o ma he-
ma ical biology 42, 347–360. doi:10.1007/s002850000075 (ci . on p. 18)
So yane, Asma, Mohammed Lahcini, Abdella i El Meziane, Mehdi Khouloud, Abdelmalek Dahchou , Syl ain Caillol,
and Mus apha Raihane (2020). P ope ies o coa ed con olled elease diammonium phospha e e ilize s p epa ed
wi h he use o bio-based amino oil. Jou nal o he Ame ican Oil Chemis s’ Socie y 97, 751–763. doi:10.1002/
aocs.12360 (ci . on p. 18)
Taylo , L, S. Banwa , P. Valdes, J. Leake, and D. Bee ling (2012). E alua ing he e ec s o he co-e olu ion o
e es ial ecosys ems, clima e and CO2on con inen al wea he ing o e geological ime: a global-scale p ocess-based
app oach. Philosophical T ansac ions o he Royal Socie y B 367, 565–582. doi:10.1098/ s b.2011.0251 (ci . on
p. 2)
Taylo , L, C D iscoll, P G o man, G Rau, J Blum, and D Bee ling (2021). Inc eased ca bon cap u e by a silica e- ea ed
o es ed wa e shed a enua ed by acid deposi ion. Biogeosciences 8, 169–188. doi:10.5194/bg-18-169-2021 (ci .
on p. 2)
Taylo , L, M Lomas, Y Kanzaki, E Bol on, N Plana sky, C Reinha d, M Kelland, and D Bee ling (2023). Rock-MIP:
The Enhanced Rock Wea he ing Model In e compa ison P ojec . Ame ican Geophysical Union abs ac GC51K-
0742, 15 Decembe 2023, San F ancisco (ci . on p. 2)
Taylo , L, J Qui k, R Tho ley, P Kha echa, J Hansen, A Ridgwell, M Lomas, S Banwa , and D Bee ling (2016).
Enhanced wea he ing s a egies o s abilizing clima e and a e ing ocean acidi ica ion. Na u e Clima e Change 6,
402–406. doi:10.1038/nclima e2882 (ci . on p. 2)
Taylo , L, C Reinha d, M Lomas, Y Kanzaki, E Bol on, N Plana sky, and D Bee ling (2024). Enhanced Rock Wea he -
ing Model In e compa ison P ojec Phase 1: Responses o Silica e Feeds ocks and soil pCO2. Ame ican Geophysical
Union abs ac 1681700 alk B14B-03, 15 Decembe 2024, Washing on DC (ci . on p. 2)
Tes e , Je e son W, W Gab iel Wo ley, B uce A Robinson, Cha les O G igsby, and Je ey L Fee e (1994). Co ela ing
qua z dissolu ion kine ics in pu e wa e om 25 o 625 ◦C. Geochimica e cosmochimica ac a 58, 2407–2420. doi:
10.1016/0016-7037(94)90020-5 (ci . on p. 33)
