Safe and just Earth system boundaries

Na u e | www.na u e.com | 1
A icle
Sa e and jus Ea h sys em bounda ies
Johan Rocks öm1,2,3 ✉, Joyee a Gup a4,5, Dahe Qin6,7,8, S e en J. Lade3,9,10 ✉, Jesse F. Ab ams11,
Lau en S. Ande sen1, Da id I. A ms ong McKay3,11,12, Xuemei Bai10, Go indasamy Bala13,
S ua E. Bunn14, Daniel Ciobanu3, Fab ice DeCle ck15,16, K is ie Ebi17, Lau en Gi o d18,
Ch is ophe Go don19, Syezlin Hasan14, No ichika Kanie20, Timo hy M. Len on11,
Sina Lo iani1, Diana M. Li e man18, Awaz Mohamed21, Nebojsa Nakiceno ic22,
Da id Obu a23, Daniel Ospina9, Klaudia P odani4, C elis Rammel 4, Bo is Sakschewski1,
Joe i Schol ens4, Ben S ewa -Kos e 14, Thejna Tha ammal24, De le an Vuu en25,26,
Pe e H. Ve bu g27,28, Rica da Winkelmann1,29, Ca oline Zimm22, Elena M. Benne 30,31,
S e an B ingezu32, Wendy B oadga e9, Pamela A. G een33, Lei Huang34, Lisa Jacobson9,
Ch is ophe Ndehedehe14,35, Simona Pedde9,36, Juan Rocha3,9, Ma en Sche e 37,
Lena Schul e-Uebbing25,38, Wim de V ies38, Cunde Xiao6,39, Chi Xu40, Xinwu Xu7,8,
Noelia Za a-Cal o41 & Xin Zhang42
The s abili y and esilience o he Ea h sys em and human well-being a e insepa ably
linked1–3, ye hei in e dependencies a e gene ally unde - ecognized; consequen ly,
hey a e o en ea ed independen ly4,5. He e, we use modelling and li e a u e
assessmen o quan i y sa e and jus Ea h sys em bounda ies (ESBs) o clima e, he
biosphe e, wa e and nu ien cycles, and ae osols a global and subglobal scales.
We p opose ESBs o main aining he esilience and s abili y o he Ea h sys em (sa e
ESBs) and minimizing exposu e o signi ican ha m o humans om Ea h sys em
change (a necessa y bu no su icien condi ion o jus ice)4. The s ic e o he sa e
o jus bounda ies se s he in eg a ed sa e and jus ESB. Ou indings show ha jus ice
conside a ions cons ain he in eg a ed ESBs mo e han sa e y conside a ions o
clima e and a mosphe ic ae osol loading. Se en o eigh globally quan i ied sa e and
jus ESBs and a leas wo egional sa e and jus ESBs in o e hal o global land a ea
a e al eady exceeded. We p opose ha ou assessmen p o ides a quan i a i e
ounda ion o sa egua ding he global commons o all people now and in o he
u u e.
Humani y is well in o he An h opocene6, he p oposed new geologi-
cal epoch whe e human p essu es ha e pu he Ea h sys em on a a-
jec o y mo ing apidly away om he s able Holocene s a e o he
pas 12,000 yea s, which is he only s a e o he Ea h sys em we ha e
e idence o being able o suppo he wo ld as we know i 7,8. These
apid changes o he Ea h sys em unde mine c i ical li e-suppo
sys ems
1,9,10
, wi h signi ican socie al impac s al eady el
1,3
, and hey
could lead o igge ing ipping poin s ha i e e sibly des abilize
he Ea h sys em
7,11,12
. These changes a e mos ly d i en by social and
economic sys ems un on unsus ainable esou ce ex ac ion and con-
sump ion. Con ibu ions o Ea h sys em change and he consequences
o i s impac s a y g ea ly among social g oups and coun ies. Gi en
hese in e dependencies be ween inclusi e human de elopmen and
a s able and esilien Ea h sys em1–3,13, an assessmen o sa e and jus
h ps://doi.o g/10.1038/s41586-023-06083-8
Recei ed: 23 June 2022
Accep ed: 14 Ap il 2023
Published online: xx xx xxxx
Open access
Check o upda es
1Po sdam Ins i u e o Clima e Impac Resea ch (PIK), Membe o he Leibniz Associa ion, Po sdam, Ge many. 2Ins i u e o En i onmen al Science and Geog aphy, Uni e si y o Po sdam,
Po sdam, Ge many. 3S ockholm Resilience Cen e, S ockholm Uni e si y, S ockholm, Sweden. 4Ams e dam Ins i u e o Social Science Resea ch, Uni e si y o Ams e dam, Ams e dam, The
Ne he lands. 5IHE Del Ins i u e o Wa e Educa ion, Del , The Ne he lands. 6S a e Key Labo a o y o C yosphe ic Science, No hwes Ins i u e o Eco-En i onmen and Resou ces, Chinese
Academy o Sciences, Lanzhou, China. 7China Me eo ological Adminis a ion, Beijing, China. 8Uni e si y o Chinese Academy o Sciences, Beijing, China. 9Fu u e Ea h Sec e a ia , S ockholm,
Sweden. 10Fenne School o En i onmen & Socie y, Aus alian Na ional Uni e si y, Canbe a, Aus alia. 11Global Sys ems Ins i u e, Uni e si y o Exe e , Exe e , UK. 12Geo esilience Analy ics,
Lea he head, UK. 13Cen e o A mosphe ic and Oceanic Sciences, Indian Ins i u e o Science, Bengalu u, India. 14Aus alian Ri e s Ins i u e, G i i h Uni e si y, B isbane, Aus alia. 15EAT, Oslo,
No way. 16Alliance o Bio e si y In e na ional and CIAT o he CGIAR, Mon pellie , F ance. 17Cen e o Heal h & he Global En i onmen , Uni e si y o Washing on, Sea le, WA, USA. 18School o
Geog aphy, De elopmen and En i onmen , Uni e si y o A izona, Tucson, AZ, USA. 19Ins i u e o En i onmen and Sani a ion S udies, Uni e si y o Ghana, Legon, Ghana. 20G adua e School o
Media and Go e nance, Keio Uni e si y, Fujisawa, Japan. 21Func ional Fo es Ecology, Uni e si ä Hambu g, Ba sbü el, Ge many. 22In e na ional Ins i u e o Applied Sys ems Analysis,
Laxenbu g, Aus ia. 23CORDIO Eas A ica, Mombasa, Kenya. 24In e disciplina y Cen e o Wa e Resea ch, Indian Ins i u e o Science, Bengalu u, India. 25Cope nicus Ins i u e o Sus ainable
De elopmen , U ech Uni e si y, U ech , The Ne he lands. 26PBL Ne he lands En i onmen al Assessmen Agency, The Hague, The Ne he lands. 27Swiss Fede al Ins i u e o Fo es , Snow and
Landscape Resea ch, Bi mensdo , Swi ze land. 28Ins i u e o En i onmen al S udies, V ije Uni e si ei Ams e dam, Ams e dam, The Ne he lands. 29Ins i u e o Physics and As onomy,
Uni e si y o Po sdam, Po sdam, Ge many. 30Biele School o En i onmen , McGill Uni e si y, Mon eal, Canada. 31Depa men o Na u al Resou ce Sciences, McGill Uni e si y, Mon eal,
Canada. 32Cen e o En i onmen al Sys ems Resea ch, Kassel Uni e si y, Kassel, Ge many. 33En i onmen al Sciences Ini ia i e, Ad anced Science Resea ch Cen e a he G adua e Cen e , Ci y
Uni e si y o New Yo k, New Yo k, NY, USA. 34Na ional Clima e Cen e , Beijing, China. 35School o En i onmen & Science, G i i h Uni e si y, Na han, Aus alia. 36Soil Geog aphy and Landscape
G oup, Wageningen Uni e si y & Resea ch, Wageningen, The Ne he lands. 37Depa men o En i onmen al Sciences, Wageningen Uni e si y & Resea ch, Wageningen, The Ne he lands.
38En i onmen al Sys ems Analysis G oup, Wageningen Uni e si y & Resea ch, Wageningen, The Ne he lands. 39S a e Key Labo a o y o Ea h Su ace P ocesses and Resou ce Ecology, Beijing
No mal Uni e si y, Beijing, China. 40School o Li e Sciences, Nanjing Uni e si y, Nanjing, China. 41Basque Cen e o Clima e Change bc3, Scien i ic Campus o he Uni e si y o he Basque
Coun y, Biscay, Spain. 42Appalachian Labo a o y, Uni e si y o Ma yland Cen e o En i onmen al Science, F os bu g, MD, USA. ✉e-mail: johan. ocks om@pik-po sdam.de; s e en.lade@
u u eea h.o g
2 | Na u e | www.na u e.com
A icle
bounda ies is equi ed ha accoun s o Ea h sys em esilience and
human well-being in an in eg a ed amewo k4,5.
We p opose a se o sa e and jus Ea h sys em bounda ies (ESBs)
o clima e, he biosphe e, esh wa e , nu ien s and ai pollu ion a
global and subglobal scales. These domains we e chosen o he ol-
lowing easons. They span he majo componen s o he Ea h sys em
(a mosphe e, hyd osphe e, geosphe e, biosphe e and c yosphe e)
and hei in e linked p ocesses (ca bon, wa e and nu ien cycles),
he ‘global commons’14 ha unde pin he plane ’s li e-suppo sys-
ems and, he eby, human well-being on Ea h; hey ha e impac s on
policy- ele an imescales; hey a e h ea ened by human ac i i ies;
and hey could a ec Ea h sys em s abili y and u u e de elopmen
globally. Ou p oposed ESBs a e based on exis ing schola ship, expe
judgemen and widely sha ed no ms, such as Agenda 2030. They a e
mean as a anspa en p oposal o u he deba e and e inemen by
schola s and wide socie y.
Fi s , we iden i y ‘sa e’ bounda ies a subglobal and global scales
o “main ain[ing] and enhanc[ing] he s abili y and esilience o he
Ea h sys em o e ime, he eby sa egua ding i s unc ions and abil-
i y o suppo humans and all o he li ing o ganisms”
4
. To de e mine
sa e bounda ies, we use assessmen s o ipping poin isks among
local and egional ipping elemen s, e idence on declines in Ea h
sys em unc ions, analyses o his o ical a iabili y and expe judge-
men . We assess he unce ain y in and con idence o hese ESBs.
Tipping elemen s a e hose componen s o p ocesses ha egula e
he unc ioning and s a e o he plane and ha show e idence o
ha ing h esholds a which small addi ional pe u ba ions can igge
sel - ein o cing changes ha unde mine Ea h sys em esilience15,16.
We do no exclusi ely ely on ipping poin s o se ing sa e ESBs, how-
e e , and he ESBs should no be in e p e ed as ep esen ing ipping
poin s. As a e e ence s a e o human li e suppo on Ea h, we use
an in e glacial Holocene-like Ea h sys em unc ioning domina ed
by balancing eedbacks ha cope wi h, bu e and dampen dis u -
bances. Me hods and Supplemen a y In o ma ion ha e de ails on
how sa e bounda ies a e de e mined.
Second, we use h ee c i e ia o assess whe he adhe ing o he sa e
ESBs could p o ec people om signi ican ha m (Box1): ‘in e species
jus ice and Ea h sys em s abili y’ (I1)17; ‘in e gene a ional jus ice’18
be ween pas and p esen gene a ions (I2a) and p esen and u u e
gene a ions (I2b); and ‘in agene a ional jus ice’ (I3) be ween coun-
ies19, communi ies and indi iduals h ough an in e sec ional lens20.
These c i e ia si wi hin a wide Ea h sys em jus ice amewo k ha
goes beyond plane a y and issue- ela ed jus ice o ake a mul i-le el
ans o ma i e jus ice app oach ocusing on ends (bounda ies and
access le els) and means
21,22
. Me hods and Supplemen a y In o ma ion
ha e mo e de ailed discussions o he jus ice app oach applied in his
pape . We de ine ha m as nega i e impac s on humans, communi ies
and coun ies om Ea h sys em change in addi ion o backg ound
a es. The mos ecen In e go e nmen al Panel on Clima e Change
(IPCC) epo iden i ies ‘se e e’ isks and ‘high’ easons o conce n
when ens o hund eds o millions o people a e exposed o changes
in clima e, such as inc eases in empe a u e and ex eme e en s
23
. In
his pape , we de ine signi ican ha m as widesp ead se e e exis en-
ial o i e e sible nega i e impac s on coun ies, communi ies and
indi iduals om Ea h sys em change, such as loss o li es, li elihoods
o incomes; displacemen ; loss o ood, wa e o nu i ional secu i y;
and ch onic disease, inju y o malnu i ion (a glossa y is in heSup-
plemen a y Me hods).
Thi d, we combine hese jus ice c i e ia wi h his o ical analy-
ses, in e na ional heal h s anda ds, Ea h sys em modelling and
expe judgemen o quan i y sa e and jus ESBs ha minimize
human exposu e o signi ican ha m (no signi ican ha m (NSH))
om Ea h sys em change. Minimizing signi ican ha m is a co ne -
s one o na ional and in e na ional law and co ec i e jus ice24,25.
We ocus on assessing he le els o Ea h sys em change leading o
widesp ead exposu e o signi ican ha m, which will lead o g ea e
impac s when ulne able popula ions a e exposed3. Me hods and
Supplemen a y In o ma ion ha e de ails on how jus bounda ies
a e de e mined. The jus (NSH) bounda ies desc ibed he e a e nec-
essa y bu no su icien condi ions o Ea h sys em jus ice, which
mus also enable access o esou ces o all
26
and dis ibu ional and
p ocedu al ai ness22. A ounda ion ha enables minimum access o
wa e , ood, ene gy and in as uc u e o all humans alongside a sa e
and jus (NSH) ESB ceiling o maximum allowed human p essu e on
biophysical domains could cons i u e a sa e and jus ‘co ido ’ o e
ime4,22 (Fig.1).
Ou assessmen builds upon and ad ances beyond p e ious esea ch
and science-based poli ical consensus, such as he Plane a y Bounda-
ies (PBs) amewo k27, doughnu economics28 and he Sus ainable
De elopmen Goals29 in he ollowing ways. (1) We de ine jus ESBs o
a oiding signi ican ha m using he same uni s as he sa e ESBs o he
same domains and p opose ha ac o s use he s ic e o he sa e and
jus bounda ies o in o m a ge se ing. The PBs iden i y only sa e bio-
physical bounda ies. The social goals ela ed o access o o ha m om
na u al esou ces adop ed in Agenda 2030, doughnu economics and
o he app oaches
28,30–32
a e no quan i ied in compa able uni s o exam-
ine only he consequences o human ac i i ies on he Ea h sys em, no
Box 1
The ‘3I’ jus ice c i e ia used o
analyse sa e ESBs
Fu he explana ion is in Gup a e al.22. Discussion o he ca ea s
ela ed o he jus ice app oach applied in his pape is inMe hods
and Supplemen a y In o ma ion.
In e species jus ice and Ea h sys em s abili y (I1)
In e species jus ice aims o p o ec humans, o he species and
ecosys ems, ejec ing human excep ionalism. In many domains,
in e species jus ice could be achie ed by main aining Ea h sys em
s abili y wi hin sa e ESBs.
In e gene a ional jus ice (I2a and I2b)
In e gene a ional jus ice examines ela ionships and obliga ions
be ween gene a ions, such as he legacy o g eenhouse gas
emissions o ecosys em des uc ion o you h and u u e people.
Achie ing in e gene a ional jus ice equi es ecognizing he
po en ial long- e m consequences o sho - e m ac ions and
associa ed ade-o s and syne gies ac oss ime. We de ine wo
ypes o in e gene a ional jus ice: (be ween pas and p esen ; I2a)
whe he ac ions o pas gene a ions ha e minimized signi ican
ha m o cu en gene a ions and (be ween p esen and u u e; I2b)
he esponsibili y o cu en gene a ions o minimize signi ican
ha m o u u e gene a ions.
In agene a ional jus ice: be ween coun ies, communi ies and
indi iduals (I3)
In agene a ional jus ice includes ela ionships be ween p esen
indi iduals, be ween s a es (in e na ional), among people o
di e en s a es (global) and be ween communi y membe s o
ci izens (communi a ian o na ionalis ). In e sec ional jus ice
conside s mul iple and o e lapping social iden i ies and ca ego ies
( o example, gende , ace, age, class and heal h) ha unde pin
inequali y, ulne abili y and he capaci y o espond. Achie ing
in agene a ional jus ice means minimizing signi ican ha m
caused by one coun y o ano he , one communi y o ano he and
one indi idual o ano he .
Na u e | www.na u e.com | 3
ha m o humans om Ea h sys em change. A icula ing sociopoli ical
no ions, such as Ea h sys em jus ice, and con e ing hei implica ions
in o biophysical uni s can enable a be e unde s anding o he space
wi hin which humans can unc ion. (2) We de ine global and subglobal
ESBs in mos domains. The PBs’ emphasis on he global scale can be
inapp op ia e o he assessmen and managemen o domains such as
he biosphe e33 and esh wa e 34–37. (3) We se bounda ies a mul iple
likelihood le els o Ea h sys em s a es. (4) Tipping elemen assess-
men s in clima e, biosphe e and o he Ea h sys em domains a e key,
al hough no exclusi e, e idence o ou ESBs. Recen PB assessmen s
ins ead emphasize isks ela ed o he depa u e om Holocene anges
o Ea h sys em a iabili y38.
Quan i ying ESBs
Fo each Ea h sys em domain, we i s quan i y sa e bounda ies o
main aining Ea h sys em esilience, wi h mul iple le els o likelihood
e lec ing unce ain y o a iabili y in he exac posi ion o he bound-
a y. Adhe ing o hese sa e bounda ies implemen s ou ‘in e species
jus ice and Ea h sys em s abili y’ c i e ion (I1 in Box1) and will sa e-
gua d u u e gene a ions agains signi ican ha m om Ea h sys em
change (in e gene a ional jus ice; I2b in Box1), bu i may no a oid
signi ican ha m o cu en gene a ions, pa icula ly ulne able popula-
ions (I2a and I3 in Box1). Hence, (1) we p opose ha some bounda ies
be made mo e s ingen o p o ec p esen gene a ions and ecosys-
ems; (2) we complemen sa e bounda ies wi h local-le el s anda ds o
p o ec p esen gene a ions and ecosys ems; and (3) i he bounda y
is likely o cause conside able di icul ies o p esen gene a ions, we
p opose ha i is complemen ed wi h policies ha accoun o dis ibu-
i e jus ice. We also assess he cu en s a e o he Ea h sys em wi h
espec o each sa e and jus ESB.
Clima e
We iden i y sa e ESBs o wa ming (Fig.1 and Table1) based on minimiz-
ing likelihoods o igge ing clima e ipping elemen s; main aining
biosphe e and c yosphe e unc ions; and accoun ing o Holocene
(<0.5–1.0 °C) and p e ious in e glacial (<1.5–2 °C) clima e a iabili y
(Supplemen a y Me hods). Some clima e ipping poin s, such as ci cu-
la ion collapse o Amazon dieback, ha e high unce ain y o low con-
idence in hei dynamics and po en ial wa ming h esholds
16
, bu he
complemen a y palaeoclima e and biosphe e analyses independen ly
suppo he sa e clima e ESB assessmen . C yosphe e unc ion includes
main aining pe ma os in he no he n high la i udes, pe manen pola
ice shee s and moun ain glacie s and minimizing sea ice loss. We ind
ha global wa ming beyond 1.0 °C abo e p e-indus ial le els, which
has al eady been exceeded9, ca ies a mode a e likelihood o igge -
ing ipping elemen s, such as he collapse o he G eenland ice shee
o localized ab up hawing o he bo eal pe ma os
16
. One-deg ee
Celsius global wa ming is consis en wi h he sa e limi p oposed in
1990
39
and he PB o 350 ppm CO
2
( e . 27). Abo e 1.5 °C o 2.0 °C wa m-
ing, he likelihood o igge ing ipping poin s inc eases o high o
e y high, espec i ely (high con idence in Ex ended Da a Table1).
Biosphe e damage and he isk o global ca bon sinks becoming ca bon
sou ces, po en ially igge ing u he clima e eedbacks, inc ease
subs an ially
40
. We conclude ha s abilizing a o below a sa e ESB o
1.5 °C wa ming a oids he mos se e e clima e impac s on humans and
o he species, ein o cing he 1.5 °C gua d ail se in he Pa is Ag eemen
on Clima e Change.
Assessmen o signi ican ha m om clima e change sugges s he
need o a s ic e jus (NSH) bounda y. A 1.0 °C global wa ming, ens
o millions o people we e exposed o we bulb empe a u e ex emes
(Fig.2), aising conce ns o in e - and in agene a ional jus ice. A 1.5 °C
Cu en Sa e Jus Sa e and jus align
Clima e
Func ional
in eg i y
Ae osols
(subglobal)
Phospho us
Ni ogen
G oundwa e
Su ace
wa e
Na u al
ecosys em
a ea
A
c
c
e
s
s
o
u
n
d
a
i
o
n
S
a
e
a
n
d
j
u
s
c
o
i
d
o
E
a
h
s
y
s
e
m
b
o
u
n
d
a
i
e
s
Fig. 1 | P oposed sa e and jus (NSH) ESBs. Visualiza ion o sa e ESBs (da k ed),
jus (NSH) ESBs (blue), cases whe e sa e and jus (NSH) bounda ies align (g een)
and cu en global s a es (Ea h icons). Radial axes a e no malized o sa e ESBs.
Headline o cen al es ima e global bounda ies (Table1) a e plo ed o suppo
compa ison wi h he cu en global s a e, bu we emphasize ha we ha e also
de ined subglobal bounda ies and mul iple likelihood le els o many domains
(Table1). Fo ae osols, howe e , we display he subglobal bounda ies o compa e
sa e and jus bounda ies. Fo ni ogen, we plo wi h a dashed blue line he
bounda y quan i ica ion o ha m om ni a e in g oundwa e while no ing ha
he jus bounda y mus also inco po a e sa e conside a ions ia eu ophica ion,
leading o a mo e s ingen sa e and jus bounda y. Minimum access o wa e ,
ood, ene gy and in as uc u e o all humans (do ed g een line) could
cons i u e he ounda ion o a sa e and jus ‘co ido ’ (g een illed a ea), bu we
do no quan i y his ounda ion he e. Al e na i e isualiza ions a e p esen ed
in Ex ended Da a Fig.1.
4 | Na u e | www.na u e.com
A icle
wa ming, mo e han 200 million people, disp opo iona ely hose
al eady ulne able, poo and ma ginalized (in agene a ional injus-
ice), could be exposed o unp eceden ed mean annual empe a u es
41
,
and mo e han 500 million could be exposed o long- e m sea-le el ise
(Fig.2 and Me hods). These numbe s o people ha med as ly exceed
he widely accep ed ‘lea e no one behind’ p inciple29 and unde mine
mos o he Sus ainable De elopmen Goals. Mo eo e , pas emissions
ha e al eady led o signi ican ha m, including ex eme wea he e en s,
loss o habi a by Indigenous communi ies in he A c ic, loss o land
a ea by low-lying s a es and sea-le el ise o educed g oundwa e
echa ge om changing glacial mel sys ems
3
. I e e sible impac s
om c yosphe e and biosphe e ipping elemen s ha a e commi ed
by an h opogenic g eenhouse gas emissions in he coming decades
bu which un old o e cen u ies o millennia also h ea en in e gene a-
ional jus ice (Supplemen a y Me hods). We conclude ha i exposu e
o ens o millions o people o signi ican ha m is o be a oided, he
jus (NSH) bounda y should be se a o below 1.0 °C. Since e u ning
wi hin his bounda y may no be achie able in he o eseeable u u e,
adap a ions and compensa ions o educe sensi i i y o ha m and
ulne abili y will be necessa y. Du ing he2022 Uni ed Na ions Clima e
Change Con e ence (COP-27), de eloping coun ies indeed ocused
ac i ely on issues o adap a ion, loss and damage.
Biosphe e
Fo he biosphe e, we iden i y sa e ESBs o wo complemen a y meas-
u es o biodi e si y: (1) he a ea o la gely in ac na u al ecosys ems
and (2) he unc ional in eg i y o all ecosys ems, including u ban and
ag icul u al ecosys ems (Table1). Main aining a eas o la gely in ac
na u al ecosys ems is necessa y o secu ing he Ea h sys em unc ions
on which all humans, o he species (I1 in Box1) and Ea h sys em s abil-
i y depend, including s ocks and lows o ca bon, wa e and nu ien s
and hal ing species ex inc ion (Ea h sys em na u e’s con ibu ion
o people (NCP) ia Ea h sys em unc ions). Based on clima e, wa e
and species conse a ion model ou comes, we p opose a sa e ESB
o 50–60% (medium con idence in Ex ended Da a Table1) o global
land su ace co e ed by la gely in ac na u al a eas o main ain Ea h
sys em NCP (Table1 and Supplemen a y Me hods). This ange uses he
cu en a ea o na u al land co e as a minimum alue while indica ing
Table 1 | P oposed sa e and jus (NSH) ESBs ( isualized in Fig.1)
Domain: s a e
a iable Rele an Ea h
sys em change Sa e ESB subglobal
(local/ egional) Sa e ESB globally
agg ega ed Jus (NSH) ESB Sa e and jus ESB Cu en global s a e
Clima e: global
mean su ace
empe a u e
change since
p e-indus ial
(1850–1900)
Clima e ipping
poin s; exceed
in e glacial
ange; biosphe e
unc ioning
Global clima e
bounda y se o
a oid egional
ipping poin s and
biome deg ada ion
Likelihood o passing
ipping poin s: low,
0.5–1.0 °C; mode a e,
>1.0 °C; high, >1.5 °C;
e y high, >2.0 °C
Exposu e o addi ional
signi ican ha m: mode a e,
0.5–1 °C; high, 1–1.5 °C; e y
high, >1.5 °C
1.0 °C a high
exposu e o
signi ican ha m
1.2 °C
Biosphe e:
na u al
ecosys em a ea
Loss o clima e,
wa e , biodi e si y
NCP
C i ical na u al
ecosys ems need
o be p ese ed o
es o ed
>50–60% na u al
ecosys em a ea
(depending on spa ial
dis ibu ion)
Align wi h sa e bounda y
plus ensu e dis ibu ional
jus ice
>50–60% (uppe
end) depending on
dis ibu ion
45–50% na u al
ecosys em a ea
Biosphe e:
unc ional
in eg i y
Loss o mul iple
local NCP >20–25% o each
1 km2 unde (semi-)
na u al ege a ion;
>50% in ulne able
landscapes; a
<10%, ew NCP
emain
100% o land a ea
sa is ies local bounda y Align wi h sa e bounda y >20–25% o each
1 km2 unde (semi-)
na u al ege a ion
One hi d (31–36%)o
human-domina edland
a ea sa is ies ESB
Wa e : su ace
wa e lows Collapse o
eshwa e
ecosys ems
<20% magni ude
mon hly su ace
low al e a ion
100% o land a ea
sa is ies local bounda y
(sums o 7,630 km3
pe yea global low
al e a ion budge )
Align wi h sa e plus Wo ld
Heal h O ganiza ion and
Uni ed Na ions En i onmen
P og amme quali y
s anda ds
Regional and
global sa e ESBs 66% o global land
a ea sa is ies ESB
annually (3,553 km3
pe yea global
al e a ions)
Wa e :
g oundwa e
le els
Collapse o
g oundwa e -
dependen
ecosys ems
Annual d awdown
does no exceed
a e age annual
echa ge
100% o land a ea
sa is ies local bounda y
(sums o 15,800 km3 pe
yea global d awdown)
Align wi h sa e plus Wo ld
Heal h O ganiza ion and
Uni ed Na ions En i onmen
P og amme quali y
s anda ds
Sa e ESB (and
ensu e eco e y) 53% o global land
a ea sa is ies ESB
(15,700 km3 pe yea
annual d awdown)
G een wa e 38
(p e ious
assessmen )
No assessed Mon hly oo -zone
soil mois u e
de ia es om
Holocene a iabili y
<10% o ice- ee land
a ea exceeds bounda y No assessed No assessed 18%
Nu ien cycles:
ni ogen Su ace wa e
and e es ial
ecosys em
eu ophica ion
<2.5 (1–4) mg N l−1
in su ace wa e ;
<5–20 kg N ha−1 pe
yea in e es ial
ecosys ems (biome
dependen )
Su plus, <61
(35–84) Tg N pe yea ;
o al inpu ,
<143 (87–189) Tg N
pe yea
Align wi h local sa e plus
d inking wa e (<11.3
(10–11.3) mg NO3–N l−1;
globally, <117 (111–117) Tg N
pe yea ) and any a ailable
ai pollu ion ( o example,
NH3) s anda ds
Local ESBs; and
global su plus,
57 (34–74) Tg N
pe yea
Su plus, 119 Tg N pe
yea ; o al inpu ,
232 Tg N pe yea
Nu ien cycles:
phospho us Su ace wa e
eu ophica ion
<50–100 mg P pe  m3
Su plus, <4.5–9 Tg P pe
yea ; mined inpu ,
<16 (8–17) Tg P pe yea
Align wi h local sa e
bounda y o a oid
eu ophica ion
Local and global
sa e ESBs Su plus, ~10 Tg P pe
yea ; mined inpu ,
~17 Tg P pe yea
A mosphe e:
ae osol loading Monsoon sys ems <0.25–0.50 AOD Annual mean
in e hemisphe ic AOD
di e ence: <0.15
Align wi h sa e plus
<15 μg pe  m3 mean annual
PM2.5; o he le els o
exposu e o signi ican ha m
in Supplemen a y Table11
<15 μg pe  m3 PM2.5
plus egional and
global sa e ESBs
0.05 annual mean
in e hemisphe ic AOD
di e ence
Na u e | www.na u e.com | 5
he need o es o e la gely in ac na u al a eas. The exac sa e bound-
a y depends s ongly on he demand o speci ic ecological unc ions
(which in u n depend, o example, on he emaining ca bon emis-
sions o be seques e ed) and on he spa ial dis ibu ion o he la gely
in ac na u al a ea ac oss eco egions and ecosys ems. S udies gene -
ally indica e ha up o 60% o he e es ial ea h su ace a ea may be
needed, wi h some ex ending up o 80% (Supplemen a y Me hods).
Na u al ecosys em a eas compa able wi h he 50–60% e es ial ESB
a e needed in he ocean o main ain ca bon seques a ion and minimize
addi ional ma ine species ex inc ion
42
. Biome-scale bounda ies may
be mo e s ingen : o example, o p o ec opical o es biomes due
o hei con ibu ion o clima e s abili y and mois u e ecycling. I
alloca ion and coo dina ion o es o a ion e o s a e less han op i
-
mal, he equi ed minimum a ea will be la ge . I hese bounda ies a e
ansg essed, ipping poin s in ol ing loss o biome-scale unc ional
in eg i y and associa ed NCP may be igge ed, including inc eases in
species ex inc ion a es.
Adhe ence o ou p oposed sa e ESB o he a ea o la gely in ac
na u al ecosys ems should minimize ha m o u u e gene a ions (I2b in
Box1) by secu ing biosphe e con ibu ions o all li e suppo h ough a
s able and esilien Ea h sys em and localized NCP p o ided by la gely
in ac na u e. Howe e , achie ing jus ice o cu en gene a ions
(I2a and I3 in Box1) may equi e a s ic e bounda y because he sa e ESB
does no accoun o he cu en une en dis ibu ion o la gely in ac
na u al ecosys ems needed o suppo local li elihoods
43
, especially in
poo o Indigenous communi ies
44,45
. Some people and coun ies may
di ec ly bene i om policies o main ain o inc ease na u al ecosys em
a ea
46
, while o he s may ace oppo uni y cos s
47
. Hence, o ensu e jus
dis ibu ion o la gely in ac na u al ecosys ems, a jus (NSH) bound-
a y may need o be se a he uppe end o he 50–60% sa e ange, as
alloca ion will be less han op imal o achie ing he unc ions he lowe
bounda y was op imized o . We emphasize ha na u al ecosys em a ea
includes all la gely in ac na u al a eas and no only hose cu en ly
equi ing conse a ion a en ion; i does no imply p o ec ion ha
excludes human habi a ion and sus ainable use.
Func ional in eg i y is he capaci y o u ban, ag icul u al o o he
human-modi ied ecosys ems o p o ide ecological unc ions and hei
con ibu ions o people a landscape scale, complemen ing he Ea h
sys em NCP p o ided by la ge-scale in ac na u al ecosys em a eas. We
analyse wha minimum amoun , quali y and dis ance o na u al habi a
and semina u al habi a a e needed o main ain local e es ial NCP
p o ision, including pollina ion, pes and disease con ol, wa e quali y
egula ion, soil p o ec ion, na u al haza ds mi iga ion and ec ea ion.
We iden i y ha a leas 20–25% di e se semina u al habi a including
na i e species in each squa e kilome e in human-modi ied lands is
needed o suppo he p o isioning o mul iple local NCP48. The exac
amoun and quali y equi ed di e based on landscape ype, clima e
and opog aphy; he amoun can ange up o 50% in some landscapes
ulne able o na u al haza ds, such as s eep slopes o highly e od-
ible soils. This bounda y applies o ine scales, cu en ly p oposed
as 1 km2, because NCP a e no ans e able ( o example, e osion o
landslide can only be a oided by na u al co e on he same slope) and
a e o en p o ided o suppo ed by non-mobile o limi ed mobili y
species ( o example, o aging anges o pollina ing o pes - egula ing
insec s a e limi ed o a ew hund ed me es). Abou wo hi ds o
human-domina edland a ea (app oxima ely 40% o o al land a ea)has
insu icien unc ional in eg i y (Supplemen a y Me hods), and la ge
a eas a e showing symp oms o esilience loss
49
, equi ing egene a i e
p ac ices o es o e local and Ea h sys em unc ions.
The sa e bounda y o unc ional in eg i y educes u u e expo-
su e o signi ican ha m (in e gene a ional jus ice). Loss o unc ional
in eg i y in ag icul u al ecosys ems and ci ies below he sa e bound-
a y would educe ood p oduc i i y, ecosys em capaci y o mi iga e
na u al haza ds, pollu ion and nu ien losses and inc ease eliance on
ha m ul pes icides and biocides and capaci y o choose al e na e land
uses (in agene a ional jus ice). The dependence on hese se ices is
50
75
100
125
150
175
1.00
0
500
1,000
1,500
2,000
0123
Exposed popula ion (millions o people)
Mean annual empe a u e We bulb empe a u e
Sea-le el ise (2100) Sea-le el ise (mul icen u y)
2.001.751.501.25
Global mean su ace empe a u e change (°C)
Fig. 2 | Exposu e o signi ican ha m om clima e change a di e en
le els o wa ming. We examine he exposu e o he 2010 global popula ion o
mean annual empe a u es abo e 29 °C (pu ple; linea i , P < 0.01), we bulb
empe a u es o 35 °C o an a e age o a leas 1 day pe yea (o ange; quad a ic
i , P < 0.01) and u u e sea-le el ise (blue; linea in e pola ion). Sea-le el ise is
calcula ed o 2100 (blue solid) and mul i-cen ennial (blue dashed; linea
in e pola ion) esponses o a gi en empe a u e s abiliza ion by 2100,
ep esen ing nea - e m impac s and long- e m equilib ia, espec i ely. The
inse shows he magni ica ion o we bulb empe a u e in he ange 1–2 °C.
Shading indica es one s.e.

6 | Na u e | www.na u e.com
A icle
o en highe in egions wi h mo e ulne able communi ies. Speci ic
in e en ions ha secu e unc ional in eg i y a e highly local and a e
bes implemen ed unde local au ho i y, knowledge and leade ship
50
,
wi h policy in e en ions o en needed o ensu e ha ma ginalized
g oups a e no u he disempowe ed bu a e gi en he space o use
hei knowledge and app oaches o pa icipa e in such p ocesses51.
Wa e
Fo esh wa e , we p opose wo spa ially de ined sa e ESBs based on
subglobal bounda ies ha can be agg ega ed o he global scale: (1)
a low al e a ion ESB o su ace wa e and (2) a d awdown ESB o
g oundwa e (Table1). Flow al e a ion in i e s is one o he key d i -
e s o eshwa e biodi e si y loss52, leading o declines in eshwa e
biodi e si y ha ou pace hose o e es ial and ma ine sys ems
53
and in la ge-scale NCP, such as coas al and inland ishe ies, on which
millions o people depend
54,55
. Local-scale low-ecology analyses a e
o en used o es ablish en i onmen al low needs o de ine sa e le els
o low al e a ion o indi idual wa e sheds
56
. These local-scale assess-
men s could p o ide he basis o spa ially explici sa e bounda ies
bu a e absen ac oss mos o he wo ld57. In hei absence, we p opose
ha a p esump i e subglobal sa e ESB o 20% al e a ion (inc ease o
dec ease) o mon hly su ace wa e lows compa ed wi h he p e ailing
na u al low egime be me in all i e s globally (medium con idence in
Ex ended Da a Table1). This ESB lea es 80% o lows unal e ed o mee
en i onmen al needs
58,59
, assuming ha equi ed wa e quali y s and-
a ds a e also me . The ESB is suppo ed by empi ical s udies showing
ha low al e a ions wi hin 20% suppo na i e ish species and low
al e a ion beyond his le el s ongly a ec s biodi e si y and ecosys em
s uc u e and unc ion60,61 (Supplemen a y Me hods has addi ional
e e ences suppo ing he use o his h eshold). The global ESB o
su ace wa e is ha 100% o all land a ea mee s he subglobal bounda y
by limi ing al e a ions o lows by 20% in all i e s in he wo ld. Mee -
ing he global ESB sums o a global al e a ion budge o 7,630 km3pe
yea (Supplemen a y Me hods; wi h high con idence in Ex ended Da a
Table1). Globally agg ega ed i e low al e a ions a e cu en ly less
han his igu e; howe e , we a e ou side he global ESB because he
subglobal sa e ESB is only me o 66% o land a ea (Table1) and less
han hal o he global popula ion (Supplemen a y Me hods). These
esul s a e consis en wi h ecen analyses o wa e sca ci y, which
highligh he challenge o mee ing en i onmen al low equi emen s
o suppo ecosys em se ices, such as ishe ies p oduc ion, while
ensu ing he e is su icien wa e o human needs57,62.
G oundwa e aqui e s con ibu e o base lows in many i e sys ems
and di ec ly sus ain we lands and e es ial ege a ion. Unsa e le els o
g oundwa e ex ac ion occu when d awdown exceeds eplenishmen
a es, impac ing g oundwa e -dependen ecosys ems and in some
ins ances, leading o land subsidence and i e e sible aqui e loss12,63,64.
Gi en he empo al na u e o g oundwa e echa ge and discha ge and
a lack o widesp ead consis en da a on his o ical aqui e le els, we
p opose ha he sa e ESB o annual g oundwa e d awdown o all
aqui e s be he a e age annual echa ge, wi h g oundwa e conside ed
sa e i d awdown is less han echa ge. The subglobal sa e ESB is me o
a gi en aqui e when local d awdown does no exceed a e age annual
echa ge. The global ESB o g oundwa e is ha he subglobal ESB is
me o all aqui e s a ound he wo ld. Fo he 2003–2016 pe iod, he
global sum o a e age annual echa ge is app oxima ely 16,000 km
3
pe yea (Table1 and Supplemen a y Me hods; wi h high con idence
in Ex ended Da a Table1). The g oundwa e ex ac ion ha may sa ely
occu wi hin his bounda y na u ally a ies ac oss he plane and, whe e
possible, should be de ined based on local-scale moni o ing, al hough
b oad ends can also be de e mined ia sa elli e emo e sensing
65
. We
es ima e ha we a e cu en ly ou side he global ESB because g ound-
wa e le els in 47% o basins a e cu en ly in decline (Table1).
Ou jus ice analysis o he sa e ESBs o su ace and g oundwa e
highligh s he challenges o (1) mul i-le el dis ibu ion, (2) wa e
insecu i y and (3) wa e quali y. The egional su ace and g oundwa e
ESBs a e gene ally in he long- e m in e es s o su ounding commu-
ni ies, as hey conse e u u e esh wa e (in e gene a ional jus ice:
I2b in Box1). Whe e deple ed aqui e s ha e al eady caused signi ican
en i onmen al impac s
66
, g oundwa e ex ac ion should u gen ly be
educed, and echa ge a eas should be p o ec ed o es o e aqui e s o
sa e le els (NSH o p esen gene a ions: I2a and I3 in Box1). Minimizing
signi ican ha m o cu en gene a ions also equi es he ollowing.
(1) Accoun ing o mul i-le el dis ibu ion indica es he alloca ion o
allowed al e a ions be ween communi ies, sec o s o na ions sha -
ing he wa e body, whe he di ec ly o indi ec ly ia i ual wa e .
This alloca ion is pa icula ly challenging whe e he sa e ESB equi es
d as ic educ ions in wa e use. (2) Minimizing exposu e o signi ican
ha m should accoun o wa e insecu i y in di e en egions o he
wo ld. Fo example, ha m associa ed wi h poo wa e sani a ion and
hygiene condi ions disp opo iona ely impac s he heal h o young
child en in low-income coun ies67, pa icula ly in Sub-Saha an A ica
and Sou h Asia
68
. (3) Minimizing exposu e o signi ican ha m implies
add essing su ace wa e quali y guidelines o human use69, no jus
an alloca ion o wa e quan i y. A a minimum, wa e needs o be sa e
o consump ion and i iga ion, meaning ha accep able s anda ds
o aecal coli o ms and salini y mus be me . We align ou jus (NSH)
ESBs o wa e wi h he sa e ESBs while no ing ha adhe ing o he
bounda ies would conside ably es ic cu en use and will equi e
policies o ensu e dis ibu i e jus ice.
These p oposed su ace and g oundwa e ESBs a e independen
o g een wa e s ocks. G een wa e s ocks a e c i ical o main aining
he a mosphe ic wa e cycle, which egula es seasonal p ecipi a ion
le els
34
; can suppo a signi ican p opo ion o global ag icul u al
p oduc ion
70
wi h less impac on aqua ic ecosys ems han blue wa e
use71; and a e closely ela ed o he biosphe e ESBs. A ecen assess-
men
38
p oposed a spa ially explici g een wa e bounda y o ensu e
hyd ological egula ion o e es ial ecosys ems, clima e and bio-
geochemical p ocesses by de ining a maximum allowed de ia ion
(d ying o we ing) o soil mois u e le els om mid-Holocene condi-
ions. The s a e a iable o g een wa e is de ined as he pe cen age
o ice- ee land a ea ha in any mon h has oo -zone soil mois u e
le els ou side he 95 h pe cen ile o he local baseline a iabili y.
The bounda y alue is se a 10%, co esponding o he median depa -
u e le el om mid-Holocene condi ions. We include his g een wa e
bounda y in ou se o sa e ESBs (Table1), bu we limi ou in e - and
in agene a ional jus ice analysis (I2 and I3 in Box1) o su ace and
g ound blue wa e .
Nu ien s
We se sa e ESBs o ag icul u al ni ogen (N) and phospho us (P)
su pluses o minimizing eu ophica ion o su ace wa e and e es-
ial ecosys ems due o uno , leaching and a mosphe ic N deposi ion
ia ammonia and ni ogen oxide emissions (Table1). We p opose sa e
global-scale ESBs o 61 (35–84) Tg N pe yea o ag icul u al ni ogen
su plus72 and 4.5–9.0 Tg P pe yea o c opland soil phospho us su -
plus73,74 (medium con idence in Ex ended Da a Table1). These ESBs a e
based on ecen pape s
72,74
calcula ing subglobal and global ag icul u al
nu ien losses, su pluses and inpu s om c i ical N and P concen a-
ions in wa e and ai beyond which eu ophica ion occu s (Me hods,
Table1 and Supplemen a y Me hods). These ESBs p ima ily ela e o
ag icul u e, which accoun s o app oxima ely 90% o an h opogenic
N/P inpu s o he Ea h sys em72,75. Ou ESBs a e based on ag icul u al
su pluses and losses
72,74
, al hough o compa ison wi h p e ious PB
quan i ica ions (Supplemen a y Me hods), we also p o ide co e-
sponding global inpu s assuming cu en N/P use e iciency. These
ecen s udies also accoun o non-ag icul u al sou ces, assuming
hey emain a cu en le els, and he edis ibu ion o nu ien s
om o e - e ilized o unde - e ilized egions (Supplemen a y
Me hods).
Na u e | www.na u e.com | 7
Ele a ed N and P concen a ions cause ha m h ough he conse-
quences o eu ophica ion on ecosys ems and hei se ices, such as
ishe y collapse, oxic compounds eleased by algal blooms
72,76
and he
heal h impac s o ai pollu ion om ammonia-de i ed ae osols77. Ha m
can also occu om d inking su ace o g oundwa e wi h ele a ed
ni a e concen a ions78 bu a a highe le el han he sa e N concen-
a ion o su ace wa e eu ophica ion. We he e o e align he jus
(NSH) ESBs o subglobal N and subglobal and global P wi h hei sa e
bounda ies, as human ha m om nu ien cycle dis up ion is p ima ily
d i en by en i onmen al deg ada ion. Accoun ing o signi ican ha m
om g oundwa e ni a e igh ens he global N bounda y sligh ly o 57
(34–74) Tg N pe yea (Supplemen a y Me hods). These ESBs should
be complemen ed by s anda ds o local ai and wa e pollu ion o N
and wa e pollu ion o P. Addi ional jus ice conside a ions include
lack o access o N and P e ilize s, which can h ea en ood secu i y
especially o low-income communi ies and coun ies
76
, and ex ac ion
o phospha e ock, which is a limi ed esou ce cu en ly unde pinning
ood p oduc ion bu exposes poo and ma ginalized communi ies o
mining was e, des oyed land and human igh s abuses76,79.
Ae osol pollu ion
Fo ae osols, we p opose a sa e ESB de ined by he in e hemisphe ic di -
e ence in ae osol op ical dep h (AOD) (Table1) based on e idence ha
a ising No h/Sou h Hemisphe e di e ence can igge egional-scale
ipping poin s and cause subs an ial ad e se e ec s on egional hyd o-
logical cycles, in addi ion o he exis ing PB o 0.25–0.50 AOD based
on egional conside a ions
27
. We conside AOD di e ences and hei
po en ial impac s a ising om na u al emissions, an h opogenic
emissions and s a osphe ic ae osol injec ion (sola geoenginee -
ing). Obse a ional da a o he Wes A ican monsoon ain all80 and
clima e modelling s udies o he Indian monsoon81 ha e iden i ied
po en ial shi s in he loca ion o he In e opical Con e gence Zone
igge ed by di e ences in sul a e AOD be ween he No he n and
Sou he n Hemisphe es
81
. Obse a ional s udies on he impac s o in e -
hemisphe ic AOD di e ence on he Indian monsoon a e lacking, bu
obse a ions based on pas olcanic e up ions and clima e modelling
s udies show ha an inc eased concen a ion o e lec ing ae osols in
one hemisphe e leads o p ecipi a ion dec easing in he same hemi-
sphe e’s opical monsoon egions while inc easing in he opposi e
hemisphe e80,82,83. Obse ed changes in he Sou h Asian monsoon
ha e well-unde s ood mechanisms (Supplemen a y In o ma ion) ha
a e consis en wi h he e ec s o in e hemisphe ic AOD di e ence84.
The olcanic e up ions o El Chichon in he 1980s (AOD di e ence o
0.07) and Ka mai (AOD di e ence o 0.08) p o ide empi ical exam-
ples80, while model-simula ed AOD di e ences o 0.1 and app oxi-
ma ely 0.2 lead o declining p ecipi a ion in opical monsoon egions85.
In e hemisphe ic AOD di e ence and i s impac on shi s in opical
p ecipi a ion a e sensi i e o he ae osol pa icle size and he la i u-
dinal and al i udinal dis ibu ion o e lec ing ae osols
86
. Conside -
ing his and he ange o hese s udies (app oxima ely 0.05–0.20 o
addi ional AOD di e ence), we assess ha hese shi s may become
dis up i e i he in e hemisphe ic AOD di e ence, cu en ly app oxi-
ma ely 0.0587 on a e age and app oxima ely 0.1 in he bo eal sp ing and
summe
87
, exceeds 0.15 (low con idence in Ex ended Da a Table1) due
o ai pollu ion
85
o geoenginee ing- ela ed ae osol asymme ies
81,85
(Supplemen a y Me hods).
Signi ican ha m o human heal h om exposu e o ae osols, such
as pa icula e ma e (PM), sugges s a mo e s ingen jus (NSH)
bounda y based on local ai pollu ion s anda ds
88
. PM and o he ae o-
sols a e associa ed wi h espi a o y illnesses and p ema u e dea hs
as well as hea p oblems and debili a ing as hma89. We selec a jus
(NSH) bounda y o 15 μg pe  m
3
mean annual exposu e o PM
2.5
o
a oid a high likelihood o signi ican ha m om ae osols (Table1
and Suppo ing In o ma ion) based on Wo ld Heal h O ganiza ion
Numbe o
bounda ies
ansg essed
0
1
2
3
4
5
6
7
Fig. 3 | Ho spo s o cu en ESB ansg essions. The numbe o subglobal
clima e ( wo local exposu e bounda ies), unc ional in eg i y, su ace wa e ,
g oundwa e , ni ogen, phospho us and ae osol sa e and jus ESBs cu en ly
ansg essed by loca ion. No mo e han se en o hese eigh me ics ha e hei
ESBs ansg essed in any one pixel. Since clima e is a globally de ined ESB,
we use we bulb empe a u es o o e 35 °C o a leas 1 day pe yea and low-
ele a ion coas al zones (<5 m) exposed o sea-le el ise as p oxies o local
clima e ansg ession while acknowledging ha he impac s o clima e change
a e a mo e di e se. We also emphasize ha exposu e o a loca ion does no
necessa ily imply esponsibili y o causing o add essing hese en i onmen al
impac s. We in i e he eade o in es iga e he consequences o di e en
bounda y alues using he code in he code a ailabili y in o ma ion.
8 | Na u e | www.na u e.com
A icle
202188 guidelines (Table1) and Eu opean Union and US En i on-
men al P o ec ion Agency ai quali y s anda ds
90,91
. Such local and
egional guidance is needed because PM
2.5
cha ac e is ics, such as
oxici y, a e highly place and sou ce speci ic. Eigh y- i e pe cen o
he wo ld popula ion is cu en ly exposed o PM
2.5
concen a ions
beyond his bounda y92, and exposu e o ambien PM2.5 is es ima ed
o cause 4.2 million dea hs annually89, wi h ulne able g oups being
a ec ed disp opo iona ely mo e while pollu ing less93. Ai pollu ion
scena ios based on globally success ul s ingen mi iga ion and pol-
lu ion con ol show educ ions in a ec ed popula ions, bu a eas o
high ai pollu ion migh emain
94
. A 15 μg pe  m
3
PM
2.5
concen a ion
ansla es
95,96
o an AOD o app oxima ely 0.17, indica ing ha he jus
(NSH) bounda y o ae osols is mo e s ingen han he sa e egional
bounda y (0.25–0.50) (Table1).
No el en i ies and o he pollu an s
We acknowledge he isks o Ea h sys em s abili y and human well-
being om o he ai and wa e pollu an s, o which he e a e al eady
well-accep ed guidelines
88
, and he eme ging h ea s om no el en i-
ies, new o ms o exis ing subs ances and modi ied li e o ms ha
a e geologically o e olu iona ily no el and could ha e la ge-scale
unwan ed geophysical o biological impac s on he Ea h sys em
27,97
.
E idence on he di e se isk po en ials o no el en i ies, such as mic o-
plas ics, ‘ o e e chemicals’, an ibio ics, adioac i e was e, hea y me -
als o o he eme ging con aminan s, o Ea h sys em unc ion and
human heal h and ood secu i y is inc easing, bu knowledge gaps
on he scale and scope o po en ial impac s emain
98
. Pe sson e al.
97
epo ed ha humani y has c ossed he PB o no el en i ies, al hough
da a limi a ions and quan i ica ion a e challenging e en o he known
no el en i ies. The di e en ia ed impac s o no el en i ies al eady
wi nessed oday ac oss di e en popula ions and he long li e imes o
hese subs ances aise clea in agene a ional and in e gene a ional
jus ice conce ns97,98.
Cu en s a e
Se en o he eigh global-scale sa e and jus ESBs ha we quan i ied
ha e al eady been c ossed (Fig.1 and Table1). T ansg ession o ESBs is
spa ially widesp ead, wi h wo o mo e sa e and jus ESBs ansg essed
o 52% o he wo ld’s land su ace, a ec ing 86% o he global popula-
ion (Fig.3). Some communi ies expe ience many ESB ansg essions,
wi h ou o mo e ESBs ansg essed o 28% o global popula ion bu
only 5% o global land su ace (Fig.3). Spa ial ho spo ansg essions
a e he e o e concen a ed in egions o highe popula ion densi y,
aising majo in agene a ional jus ice conce ns.
Towa d a sa e and jus u u e
We de ined and quan i ied sa e and jus (NSH) ESBs o sus aining
he global commons ha egula e he s a e o he plane , p o ec
o he species, gene a e NCP, educe signi ican ha m o humans and
suppo inclusi e human de elopmen (Fig.1 and Table1). Because
exceeding sa e bounda ies esul s in widesp ead signi ican ha m, ou
jus and sa e ESBs align o su ace wa e , g oundwa e , unc ional
in eg i y, na u al ecosys em a ea, phospho us and ni ogen. Mee ing
hese bounda ies wi hou ans o ma ion, howe e , could signi i-
can ly ha m cu en gene a ions. In wo cases, ae osols and clima e,
he jus bounda ies a e mo e s ingen han he sa e bounda ies,
which indica es ha people expe ience signi ican ha m be o e ha
Ea h sys em domain is des abilized.
We iden i ied subglobal ESBs, which, in many domains, a e he
ele an scale o ac ion o a oid loss o Ea h sys em s abili y and
minimize exposu e o signi ican ha m, and global ESBs, which a e e e -
ence poin s o moni o ing human impac s a he Ea h sys em scale.
Na ions, ci ies, businesses and o he key ac o s need o se and achie e
science-based a ge s o educing hei en i onmen al impac s based
on ansla ion o he sa e and jus ESBs o ac o ai sha es
99
. Clima e
is he only ESB ha has a ela i ely well-es ablished and implemen ed
me hodology100,101, wi h me hodologies o o he domains unde de el-
opmen 101,102. We emphasize ha ou ESBs complemen , no o e - ide,
en i onmen al es ic ions o speci ic local se ings: o example,
s ic e biosphe e bounda ies o ca bon-dense ecosys ems o a ge ed
conse a ion e o s o p o ec ing endange ed o emblema ic spe-
cies. We also acknowledge ha o he ac o s may choose o implemen
a ge s based on o he likelihood le els han hose we ha e highligh ed
(Fig.1 and Table1): o example, a lowe isk ole ance han he high
isk o passing ipping poin s associa ed wi h a 1.5 °C sa e bounda y.
We o e ou ESBs as an in eg a ion o social and na u al sciences o
u he e inemen , in he spi i ha he PBs we e p oposed o e a decade
ago
103
. Se en o he eigh globally quan i ied ESBs ha e been c ossed
and a leas wo local ESBs in much o he wo ld ha e been c ossed,
pu ing human li elihoods o cu en and u u e gene a ions a isk.
No hing less han a jus global ans o ma ion ac oss all ESBs is equi ed
o ensu e human well-being. Such ans o ma ions mus be sys emic
ac oss ene gy, ood, u ban and o he sec o s, add essing he economic,
echnological, poli ical and o he d i e s o Ea h sys em change, and
ensu e access o he poo h ough educ ions and ealloca ion o
esou ce use. All e idence sugges s his will no be a linea jou ney; i
equi es a leap in ou unde s anding o how jus ice, economics, echnol-
ogy and global coope a ion can be u he ed in he se ice o a sa e and
jus u u e.
Online con en
Any me hods, addi ional e e ences, Na u e Po olio epo ing summa-
ies, sou ce da a, ex ended da a, supplemen a y in o ma ion, acknowl-
edgemen s, pee e iew in o ma ion; de ails o au ho con ibu ions
and compe ing in e es s; and s a emen s o da a and code a ailabili y
a e a ailable a h ps://doi.o g/10.1038/s41586-023-06083-8.
1. IPBES. Global assessmen epo on biodi e si y and ecosys em se ices o he
In e go e nmen al Science-Policy Pla o m on Biodi e si y and Ecosys em Se ices.
Zenodo h ps://doi.o g/10.5281/zenodo.5657041 (2019).
2. Folke, C. e al. Ou u u e in he An h opocene biosphe e. Ambio 50, 834–869 (2021).
3. IPCC Clima e Change 2022: Impac s, Adap a ion, and Vulne abili y (eds Pö ne , H.-O.
e al.) (Camb idge Uni . P ess, 2022).
4. Rocks öm, J. e al. Iden i ying a sa e and jus co ido o people and he plane . Ea h’s
Fu u e 9, e2020EF001866 (2021).
5. Rocks öm, J. e al. S ockholm o S ockholm: achie ing a sa e Ea h equi es goals ha
inco po a e a jus app oach. One Ea h 4, 1209–1211 (2021).
6. Zalasiewicz, J. e al. The Wo king G oup on he An h opocene: summa y o e idence and
in e im ecommenda ions. An h opocene 19, 55–60 (2017).
7. S e en, W. e al. T ajec o ies o he Ea h sys em in he An h opocene. P oc. Na l Acad.
Sci. USA 115, 8252–8259 (2018).
8. Xu, C., Kohle , T. A., Len on, T. M., S enning, J.-C. & Sche e , M. Fu u e o he human
clima e niche. P oc. Na l Acad. Sci. USA 117, 11350–11355 (2020).
9. IPCC Clima e Change 2021: The Physical Science Basis (eds Masson-Delmo e, V. e al.)
(Camb idge Uni . P ess, 2021).
10. UNEP Global En i onmen Ou look—GEO-6: Heal hy Plane , Heal hy People (Camb idge
Uni . P ess, 2019); h ps://doi.o g/10.1017/9781108627146.
11. Len on, T. M. e al. Clima e ipping poin s— oo isky o be agains . Na u e 575, 592–595
(2019).
12. UNEP Global En i onmen Ou look—GEO-6: Technical Summa y (Camb idge Uni . P ess,
2021); h ps://wedocs.unep.o g/20.500.11822/32024.
13. Bie mann, F., Di h, E. & Kal agianni, A. Plane a y jus ice as a challenge o ea h sys em
go e nance: edi o ial. Ea h Sys em Go e nance 6, 100085 (2020).
14. Nakiceno ic, N., Rocks öm, J., Ga ney, O. & Zimm, C. Global Commons in he
An h opocene: Wo ld De elopmen on a S able and Resilien Plane . IIASA Wo king Pape
(IIASA, 2016); h p://pu e.iiasa.ac.a /14003/.
15. Len on, T. M. e al. Tipping elemen s in he Ea h’s clima e sys em. P oc. Na l Acad. Sci.
USA 105, 1786–1793 (2008).
16. A ms ong McKay, D. I. e al. Exceeding 1.5 °C global wa ming could igge mul iple
clima e ipping poin s. Science 377, eabn7950 (2022).
17. Bu ke, A. & Fishel, S. in Non-Human Na u e in Wo ld Poli ics: Theo y and P ac ice (eds
Pe ei a, J. C. & Sa amago, A.) 33–52 (Sp inge In e na ional Publishing, 2020).
18. Meye , L. In e gene a ional jus ice. In The S an o d Encyclopedia o Philosophy (ed. Zal a,
E. N.) (S an o d, 2021); h ps://pla o.s an o d.edu/a chi es/sum2021/en ies/jus ice-
in e gene a ional/.
19. Blake, M. & Smi h, P. T. In e na ional dis ibu i e jus ice. In The S an o d Encyclopedia o
Philosophy (ed. Zal a, E. N.) (S an o d, 2022); h ps://pla o.s an o d.edu/a chi es/
sum2022/en ies/in e na ional-jus ice/.
Na u e | www.na u e.com | 9
20. No lock, K. Feminis e hics. In The S an o d Encyclopedia o Philosophy (ed. Zal a, E. N.)
(S an o d, 2019); h ps://pla o.s an o d.edu/a chi es/sum2019/en ies/ eminism-e hics/.
21. Gup a, J. e al. Reconciling sa e plane a y a ge s and plane a y jus ice: why should social
scien is s engage wi h plane a y a ge s? Ea h Sys em Go e nance 10, 100122 (2021).
22. Gup a, J. e al. Ea h sys em jus ice needed o iden i y and li e wi hin Ea h sys em
bounda ies. Na . Sus ain. h ps://doi.o g/10.1038/s41893-023-01064-1 (2023).
23. O’Neill, B. e al. in Clima e Change 2022: Impac s, Adap a ion, and Vulne abili y (eds
Pö ne , H.-O. e al.) 2411–2538 (Camb idge Uni . P ess, 2022).
24. Gup a, J. & Schmeie , S. Fu u e p oo ing he p inciple o no signi ican ha m. In . En i on.
Ag eem. 20, 731–747 (2020).
25. Spijke s, O. The no signi ican ha m p inciple and he human igh o wa e . In . En i on.
Ag eem. 20, 699–712 (2020).
26. Rammel , C. e al. Impac s o mee ing minimum access on c i ical ea h sys ems amids
he G ea Inequali y. Na . Sus ain. 6, 212–221 (2022).
27. S e en, W. e al. Plane a y bounda ies: guiding human de elopmen on a changing
plane . Science 347, 1259855 (2015).
28. Rawo h, K. A doughnu o he An h opocene: humani y’s compass in he 21s cen u y.
Lance Plane Heal h 1, e48–e49 (2017).
29. UN GA. T ans o ming Ou Wo ld: The 2030 Agenda o Sus ainable De elopmen Gene al
Assembly esolu ion 70/1 ol. A/RES/70/1 (Uni ed Na ions, 2015).
30. an Vuu en, D. P. e al. De ining a sus ainable de elopmen a ge space o 2030 and
2050. One Ea h 5, 142–156 (2022).
31. Hickel, J. Is i possible o achie e a good li e o all wi hin plane a y bounda ies? Thi d
Wo ld Q. 40, 18–35 (2019).
32. O’Neill, D. W., Fanning, A. L., Lamb, W. F. & S einbe ge , J. K. A good li e o all wi hin
plane a y bounda ies. Na . Sus ain. 1, 88–95 (2018).
33. Mace, G. M. e al. App oaches o de ining a plane a y bounda y o biodi e si y. Glob.
En i on. Change 28, 289–297 (2014).
34. Gleeson, T. e al. The wa e plane a y bounda y: in e oga ion and e ision. One Ea h 2,
223–234 (2020).
35. Zippe , S. C. e al. In eg a ing he wa e plane a y bounda y wi h wa e managemen om
local o global scales. Ea h’s Fu u e 8, e2019EF001377 (2020).
36. Heis e mann, M. HESS opinions: a plane a y bounda y on eshwa e use is misleading.
Hyd ol. Ea h Sys . Sci. 21, 3455–3461 (2017).
37. Bie mann, F. & Kim, R. E. The bounda ies o he plane a y bounda y amewo k: a c i ical
app aisal o app oaches o de ine a ‘sa e ope a ing space’ o humani y. Annu. Re .
En i on. Resou . 45, 497–521 (2020).
38. Wang-E landsson, L. e al. A plane a y bounda y o g een wa e . Na . Re . Ea h En i on.
3, 380–392 (2022).
39. Rijsbe man, F. R. & Swa , R. J. (eds) Ta ge s and Indica o s o Clima e Change. Repo o
Wo king G oup II o he Ad iso y G oup on G eenhouse Gases (S ockholm En i onmen al
Ins i u e, 1990).
40. Pa mesan, C. e al. in Clima e Change 2022: Impac s, Adap a ion, and Vulne abili y (eds
Pö ne , H.-O. e al.) 197–377 (Camb idge Uni . P ess, 2022).
41. Len on, T. M. e al. Quan i ying he human cos o global wa ming. Na . Sus ain.h ps://
doi.o g/10.1038/s41893-023-01132-6(2023).
42. Sala, E. e al. P o ec ing he global ocean o biodi e si y, ood and clima e. Na u e 592,
397–402 (2021).
43. Fedele, G., Dona i, C. I., Bo nacelly, I. & Hole, D. G. Na u e-dependen people: mapping
human di ec use o na u e o basic needs ac oss he opics. Glob. En i on. Change 71,
102368 (2021).
44. Vi a, B. & Kon oleon, A. in Biodi e si y Conse a ion and Po e y Alle ia ion: Explo ing he
E idence o a Link (eds Roe, D. e al.) 52–84 (Wiley, 2012).
45. Al es, R. R. N. & Rosa, I. M. L. Biodi e si y, adi ional medicine and public heal h: whe e
do hey mee ? J. E hnobiol. E hnomed. 3, 14 (2007).
46. Isbell, F. e al. Linking he in luence and dependence o people on biodi e si y ac oss
scales. Na u e 546, 65–72 (2017).
47. Ellis, E. C. & Meh abi, Z. Hal Ea h: p omises, pi alls, and p ospec s o dedica ing hal o
Ea h’s land o conse a ion. Cu . Opin. En i on. Sus ain. 38, 22–30 (2019).
48. Ga ibaldi, L. A. e al. Wo king landscapes need a leas 20% na i e habi a . Conse . Le .
14, e12773 (2020).
49. Rocha, J. C. Ecosys ems a e showing symp oms o esilience loss. En i on. Res. Le . 17,
065013 (2022).
50. Obu a, D. O. e al. In eg a e biodi e si y a ge s om local o global le els. Science 373,
746–748 (2021).
51. Pascual, U. e al. Biodi e si y and he challenge o plu alism. Na . Sus ain. 4, 567–572
(2021).
52. Tickne , D. e al. Bending he cu e o global eshwa e biodi e si y loss: an eme gency
eco e y plan. Bioscience 70, 330–342 (2020).
53. Reid, A. J. e al. Eme ging h ea s and pe sis en conse a ion challenges o eshwa e
biodi e si y. Biol. Re . Camb. Philos. Soc. 94, 849–873 (2019).
54. Dodds, W. K., Pe kin, J. S. & Ge ken, J. E. Human impac on eshwa e ecosys em se ices:
a global pe spec i e. En i on. Sci. Technol. 47, 9061–9068 (2013).
55. Funge-Smi h, S. & Benne , A. A esh look a inland ishe ies and hei ole in ood
secu i y and li elihoods. Fish Fish 20, 1176–1195 (2019).
56. Po , N. L. e al. The ecological limi s o hyd ologic al e a ion (ELOHA): a new amewo k
o de eloping egional en i onmen al low s anda ds. F eshw. Biol. 55, 147–170 (2010).
57. Liu, X. e al. En i onmen al low equi emen s la gely eshape global su ace wa e
sca ci y assessmen . En i on. Res. Le . 16, 104029 (2021).
58. Hoeks a, A. Y., Mekonnen, M. M., Chapagain, A. K., Ma hews, R. E. & Rich e , B. D. Global
mon hly wa e sca ci y: blue wa e oo p in s e sus blue wa e a ailabili y. PLoS ONE 7,
e32688 (2012).
59. Rich e , B. D., Da is, M. M., Apse, C. & Kon ad, C. A p esump i e s anda d o en i onmen al
low p o ec ion. Ri e Res. Appl. 28, 1312–1321 (2012).
60. Rolls, R. J. & A hing on, A. H. How do low magni udes o hyd ologic al e a ion impac
i e ine ish popula ions and assemblage cha ac e is ics? Ecol. Indic. 39, 179–188
(2014).
61. Ca lisle, D. M., Wolock, D. M. & Meado , M. R. Al e a ion o s eam low magni udes and
po en ial ecological consequences: a mul i egional assessmen . F on . Ecol. En i on. 9,
264–270 (2010).
62. Mekonnen, M. M. & Hoeks a, A. Y. Fou billion people acing se e e wa e sca ci y. Sci. Ad .
2, e1500323 (2016).
63. Minde houd, P. S. J., Middelkoop, H., E kens, G. & S ou hame , E. G oundwa e ex ac ion
may d own mega-del a: p ojec ions o ex ac ion-induced subsidence and ele a ion o
he Mekong del a o he 21s cen u y. En i on. Res. Commun. 2, 011005 (2020).
64. Ka h, J., Boul on, A. J., Ha ison, E. T. & Dye , F. J. A concep ual amewo k o ecological
esponses o g oundwa e egime al e a ion (FERGRA). Ecohyd ol. 11, e2010 (2018).
65. Döll, P., F i sche, M., Eicke , A. & Mülle Schmied, H. Seasonal wa e s o age a ia ions as
impac ed by wa e abs ac ions: compa ing he ou pu o a global hyd ological model
wi h GRACE and GPS obse a ions. Su . Geophys. 35, 1311–1331 (2014).
66. Scanlon, B. R. e al. G oundwa e deple ion and sus ainabili y o i iga ion in he US High
Plains and Cen al Valley. P oc. Na l Acad. Sci. USA 109, 9320–9325 (2012).
67. P üss-Us ün, A. e al. Bu den o disease om inadequa e wa e , sani a ion and hygiene o
selec ed ad e se heal h ou comes: an upda ed analysis wi h a ocus on low- and
middle-income coun ies. In . J. Hyg. En i on. Heal h 222, 765–777 (2019).
68. UNESCO WWAP The Uni ed Na ions Wo ld Wa e De elopmen Repo 3: Wa e in a
Changing Wo ld (UNESCO and Ea hscan, 2009); h ps://unesdoc.unesco.o g/a k:/48223/
p 0000181993.
69. WHO Guidelines o D inking-wa e Quali y 4 h edn (Wo ld Heal h O ganiza ion, 2022);
h ps://www.who.in /publica ions/i/i em/9789240045064.
70. Rocks öm, J., Lanne s ad, M. & Falkenma k, M. Assessing he wa e challenge o a new
g een e olu ion in de eloping coun ies. P oc. Na l Acad. Sci. USA 104, 6253–6260
(2007).
71. Aldaya, M. M., Allan, J. A. & Hoeks a, A. Y. S a egic impo ance o g een wa e in
in e na ional c op ade. Ecol. Econ. 69, 887–894 (2010).
72. Schul e-Uebbing, L. F., Beusen, A. H. W., Bouwman, A. F. & de V ies, W. F om plane a y
o egional bounda ies o ag icul u al ni ogen pollu ion. Na u e 610, 507–512 (2022).
73. Zhang, X. e al. Quan i a i e assessmen o ag icul u al sus ainabili y e eals di e gen
p io i ies among na ions. One Ea h 4, 1262–1277 (2021).
74. Sp ingmann, M. e al. Op ions o keeping he ood sys em wi hin en i onmen al limi s.
Na u e 562, 519–525 (2018).
75. Zhang, X. e al. Quan i ying nu ien budge s o sus ainable nu ien managemen . Glob.
Biogeochem. Cycles 34, e2018GB006060 (2020).
76. Co dell, D. & Whi e, S. Li e’s bo leneck: sus aining he wo ld’s phospho us o a ood
secu e u u e. Annu. Re . En i on. Resou . 39, 161–188 (2014).
77. Gu, B. e al. Aba ing ammonia is mo e cos -e ec i e han ni ogen oxides o mi iga ing
PM2.5 ai pollu ion. Science 374, 758–762 (2021).
78. Wa d, M. H. e al. D inking wa e ni a e and human heal h: an upda ed e iew. In . J.
En i on. Res. Public Heal h 15, 1557 (2018).
79. Ti ado, R. & Allsopp, M. Phospho us in Ag icul u e: P oblems and Solu ions. Technical
epo ( e iew) (G eenpeace, 2012); h ps://www.g eenpeace. o/g eenpeace/wp-con en /
uploads/2012/06/ i ado-and-allsopp-2012-phospho us-in-ag icul u e- echnical- epo -
02-2012.pd .
80. Haywood, J. M., Jones, A., Bellouin, N. & S ephenson, D. Asymme ic o cing om
s a osphe ic ae osols impac s Sahelian ain all. Na . Clim. Change 3, 660–665
(2013).
81. K ishnamohan, K. S. & Bala, G. Sensi i i y o opical monsoon p ecipi a ion o he la i ude
o s a osphe ic ae osol injec ions. Clim. Dyn. 59, 151–168 (2022).
82. Liu, F. e al. Global monsoon p ecipi a ion esponses o la ge olcanic e up ions. Sci. Rep.
6, 24331 (2016).
83. Zuo, M., Zhou, T. & Man, W. Hyd oclima e esponses o e global monsoon egions
ollowing olcanic e up ions a di e en la i udes. J. Clim. 32, 4367–4385 (2019).
84. Dou ille, H. e al. in Clima e Change 2021: The Physical Science Basis (eds Masson-
Delmo e, V. e al.) 1055–1210 (Camb idge Uni . P ess, 2021).
85. Visioni, D. e al. Seasonally modula ed s a osphe ic ae osol geoenginee ing al e s he
clima e ou comes. Geophys. Res. Le . 47, e2020GL088337 (2020).
86. Zhao, M., Cao, L., Bala, G. & Duan, L. Clima e esponse o la i udinal and al i udinal
dis ibu ion o s a osphe ic sul a e ae osols. J. Geophys. Res. 126, e2021JD035379
(2021).
87. Vogel, A. e al. Unce ain y in ae osol op ical dep h om mode n ae osol‐clima e models,
eanalyses, and sa elli e p oduc s.J. Geophys. Res. 127, e2021JD035483 (2022).
88. WHO Global Ai Quali y Guidelines: Pa icula e Ma e (PM2.5 and PM10), Ozone, Ni ogen
Dioxide, Sul u Dioxide and Ca bon Monoxide (WHO, 2021); h ps://apps.who.in /i is/
handle/10665/345329.
89. Cohen, A. J. e al. Es ima es and 25-yea ends o he global bu den o disease
a ibu able o ambien ai pollu ion: an analysis o da a om he Global Bu den o
Diseases S udy 2015. Lance 389, 1907–1918 (2017).
90. EPA. Re iew o he na ional ambien ai quali y s anda ds o pa icula e ma e .
En i onmen al P o ec ion Agency. 40 CFR Pa 50. Fed. Regis. Rules Regul. 85,
82684–82748 (2020).
91. Eu opean Commission. Ai quali y s anda ds h ps://ec.eu opa.eu/en i onmen /ai /
quali y/s anda ds.h m (2020).
92. Shaddick, G. e al. Da a in eg a ion o he assessmen o popula ion exposu e o
ambien ai pollu ion o global bu den o disease assessmen . En i on. Sci. Technol. 52,
9069–9078 (2018).
93. Rao, N. D., Kiesewe e , G., Min, J., Pachau i, S. & Wagne , F. Household con ibu ions o
and impac s om ai pollu ion in India. Na . Sus ain. 4, 859–867 (2021).
94. Rao, S. e al. Fu u e ai pollu ion in he sha ed socio-economic pa hways. Glob. En i on.
Change 42, 346–358 (2017).
95. an Donkelaa , A., Ma in, R. V. & Pa k, R. J. Es ima ing g ound-le el PM2.5 using ae osol
op ical dep h de e mined om sa elli e emo e sensing. J. Geophys. Res. 111, D21201
(2006).
96. Gup a, P. e al. Sa elli e emo e sensing o pa icula e ma e and ai quali y assessmen
o e global ci ies. A mos. En i on. 40, 5880–5892 (2006).
A icle
Ex ended Da a Table 1 | Assessmen o le els o con idence in each domain’s sa e Ea h sys em bounda ies
Fo mo e in o ma ion seeMe hods. The obus ness o e idence and deg ee o ag eemen o all ESB quan i ica ions a e based on he assessmen o a ailable li e a u e and wo king g oup
expe s’ iews.