Resea ch Pape
New Yo k Ci y 2100: En i onmen al jus ice implica ions o u u e scena ios
o add essing ex eme hea
Maya Du a
a
, Pablo He e os-Can is
b,c,d
, Timon McPhea son
d,e,
, Ahmed Mus a a
d
,
Ma hew I. Palme
a
, Mika Tosca
g
, Jenni e Ven ella
d
, Elizabe h M. Cook
h,*
a
Depa men o Ecology, E olu ion, and En i onmen al Biology, Columbia Uni e si y, New Yo k, NY, USA
b
Basque Cen e o Clima e Change (BC3), 48940, Leioa, Spain
c
Ins i u e o En i onmen al Science and Technology (ICTA), Uni e si a Au `
onoma de Ba celona (UAB), Ce danyola del Vall`
es, Spain
d
U ban Sys ems Lab, The New School, New Yo k, NY, USA
e
Ca y Ins i u e o Ecosys em S udies, Millb ook, NY, USA
S ockholm Resilience Cen e, S ockholm Uni e si y, and Beije Ins i u e o Ecological Economics, Royal Swedish Academy o Sciences, S ockholm, Sweden
g
Sola punk P ojec , Chicago, IL, USA
h
En i onmen al Science Depa men , Ba na d College, 3009 B oadway, New Yo k, NY, USA
HIGHLIGHTS
•NYC s akeholde s co-p oduced u u e scena ios o u ban esilience o ex eme hea .
•Business-as-usual land de elopmen ein o ces exis ing en i onmen al injus ices.
•Co-p oduced g een in as uc u e scena io inc eased hea mi iga ion by 108%
•Co-p oduced g een in as uc u e scena io inc eased lood mi iga ion by 55%
•An icipa o y u ban planning can add ess clima e esilience and en i onmen al jus ice.
ARTICLE INFO
Keywo ds:
En i onmen al jus ice
Clima e esilience
Scena io planning
U ban ecosys em se ices
ABSTRACT
Clima e-d i en haza ds, such as ex eme hea o p ecipi a ion, a e h ea ening he cu en and u u e li abili y o
New Yo k Ci y (NYC) and disp opo iona ely a ec ing low-income communi ies and communi ies o colo . To
en ision u u e clima e esilience, go e nmen s akeholde s and esea che s co-p oduced u u e scena ios o
2100 in esponse o clima e haza ds o NYC du ing pa icipa o y wo kshops in Fall 2021. A commonly co-
p oduced s a egy included u ban g een in as uc u e (UGI) because o i s po en ial o e ain uno and p o-
ide cooling bene i s. We ask, wha a e he po en ial en i onmen al jus ice implica ions o ecosys em se ices
p o isioned om UGI dis ibu ion in he co-p oduced NYC u u e scena io compa ed o a business-as-usual u u e
scena io? To analyze po en ial ou comes and adeo s, we in eg a ed spa ially-explici UGI s a egies in o
simula ed land use and co e models. We hen assessed wo ecosys em se ices ( lood and hea mi iga ion) using
he spa ially-explici ool In eg a ed Valua ion o Ecosys em Se ices and T adeo s (InVEST). We explo ed po-
en ial en i onmen al jus ice implica ions by compa ing he p o ision o ecosys em se ices o sociodemog aphic
indica o s wi hin census block g oups be ween scena ios. P esen ly, ecosys em se ices a e disp opo iona ely
lowe o communi ies o colo , including p edominan ly Asian, Black/A ican-Ame ican, and Hispanic/La ino
communi ies. In u u e scena ios we ound ecosys em se ice p o ision will dec ease wi hin hese communi ies
unde business-as-usual land de elopmen . The u u e scena io co-p oduced o ex eme hea esilience, how-
e e , shows an inc ease in o e all p o isioning ac oss NYC, including in neighbo hoods wi h a high p opo ion o
people o colo . Ou esul s show ha co-p oduced u u e scena ios can be used o in o m s a egic u u e
planning o inclusi e adap a ion decisions o imp o e u u e clima e esilience and jus ice.
* Co esponding au ho .
E-mail add ess: [email p o ec ed] (E.M. Cook).
Con en s lis s a ailable a ScienceDi ec
Landscape and U ban Planning
jou nal homepage: www.else ie .com/loca e/landu bplan
h ps://doi.o g/10.1016/j.landu bplan.2024.105249
Recei ed 4 Feb ua y 2024; Recei ed in e ised o m 29 Oc obe 2024; Accep ed 30 Oc obe 2024
Landscape and U ban Planning 254 (2025) 105249
A ailable online 18 No embe 2024
0169-2046/© 2024 The Au ho (s). Published by Else ie B.V. This is an open access a icle unde he CC BY-NC license (
h p://c ea i ecommons.o g/licenses/by-
nc/4.0/ ).
1. In oduc ion
1.1. Clima e change, en i onmen al jus ice, and he u u e o ci ies
Clima e-d i en haza ds h ea en he cu en and u u e li abili y o
ci ies, wi h ma ginalized communi ies acing disp opo iona e isk and
ulne abili y (Dodman e al. 2022; Pelling &Ga schagen 2019). The
esilience o ci ies o clima e haza ds—such as ex eme hea o p ecip-
i a ion—is uniquely challenged by densi y and di e si y o people,
in as uc u e, and landscapes in e ac ing wi h legacies o unjus land
use (F an zeskaki e al. 2019; H¨
olsche e al. 2019). In esponse, ci ies
a e implemen ing s a egies o adap o ex eme clima e e en s (Chu &
Cannon 2021). The e icacy, howe e , depends on a ci y’s abili y o
in eg a e equi y and unce ain y in o planning gi en u ban a eas ace a
ange o plausible u u es wi h a ying implica ions o clima e esil-
ience and jus ice (Balk e al. 2022; G abowski e al. 2023).
Clima e p ojec ions indica e a high likelihood o inc eased e-
quency, in ensi y, and du a ion o wea he - ela ed haza ds (Dodman
e al. 2022; Solecki &Rosenzweig 2019). Impac s o clima e haza d-
s— anging om heal h and inancial bu dens o he loss o li es and
homes—a e no expe ienced equi ably (NYCEM 2019; Ra cli e e al.
2020). In U.S. ci ies, u ban de elopmen oo ed in acial-capi alism and
se le -colonialism, unde gi ding p ac ices including edlining, zoning,
and public housing dis ibu ion (Ro hs ein 2017), has p oduced unjus
ci ies. F om an en i onmen al jus ice lens, unjus ci ies, in pa , en ail a
disp opo iona ely high exposu e o en i onmen al haza ds ye simul-
aneously lowe alloca ion, unc ion, and quali y o g eenspaces o
communi ies o colo and low-income communi ies (Hoo e &Lim
2021; Nyelele &K oll 2020; Schell e al. 2020; Taylo , 2014). En i-
onmen al jus ice in ci ies is uniquely shaped, ampli ied, and challenged
by densi y, spa ially a ied land use, and limi ed a o dable esou ces
(Wamsle , B ink, &Ri e a 2013). Despi e hese legacies, planning e -
o s do no sys ema ically explo e po en ial u u e jus ice implica ions
o clima e planning decisions (Anguelo ski 2016).
To equi ably add ess he impac s o u u e haza ds, ci ies mus ake
an in eg a i e, an icipa o y app oach o long- e m clima e esilience
and jus ice planning (Chu &Cannon 2021; Mu˜
noz-E ickson e al. 2021).
An icipa o y planning p o ides ci ies he oppo uni y o explo e u u e
scena ios—plausible and cohe en u u e isions— ha in eg a e sys-
ems hinking h ough co-p oduc ion. Co-p oduc ion en ails a p ocess
o di e se s akeholde s wi h plu alis ic alues—o en including scien-
is s, u ban designe s, decision-make s, and/o communi y membe -
s— o engage in planning, en ision wha u u e scena ios look like and
how o implemen hem (Iwaniec e al. 2020a,2020b; Pe ei a e al.
2018). Fu u e-o ien ed planning is a ool o mo e pas pe cei ed p esen
cons ain s, unce ain cycles o poli ical will and “business-as-usual”
scena ios, whe e ci ies con inue on cu en ajec o ies wi hou in en-
ional shi s owa ds no ma i e o adically ans o ma i e goals o
alues (Mu˜
noz-E ickson e al. 2021). In an e o o add ess injus ices,
an icipa o y planning can conside u u e unknowns, iden i y pa hways
o each desi ed ou comes, and explo e po en ial ou comes and unin-
ended adeo s o decisions o allow o adjus men (Be b´
es-Bl´
azquez
e al. 2023).
1.2. U ban g een in as uc u e and ecosys em se ices
Gi en he complex challenges clima e change b ings o ci ies, man-
agemen o en inco po a es s a egies wi h po en ially wide- eaching
bene i s (F an zeskaki e al. 2019; H¨
olsche e al. 2019). One inc eas-
ingly popula mul i unc ional s a egy o enhance clima e esilience
includes u ban g een in as uc u e (UGI) (G abowski e al. 2022;
G´
omez-Bagge hun &Ba on 2013; Hansen e al. 2015). UGI in eg a es
ecological and echnological-enginee ed in as uc u e (G abowski e al.
2023) h ough he es o a ion, connec ion, and/o design o g een space
o p o ide bene i s o human well-being (He e os-Can is &McPhea son
2021; McPhea son 2022). UGI is concep ualized in di e se ways and,
depending on desi ed ou comes, can consis o g een oo s, ee can-
opies, and pa ks wi h di e se ea u es (F ancis and Jensen, 2017; Lan-
gemeye e al. 2015; Ma sle e al., 2021; Nyelele e al. 2019). UGI is
pa icula ly appealing due o i s abili y o p o ide bene i s o people,
e e ed o as ecosys em se ices. Fo example, uno e en ion and
cooling can be inc eased h ough he implemen a ion o UGI. Ecosys em
se ices ha e become an impo an amewo k o plan and measu e he
clima e esilien bene i s o UGI (Genele i e al. 2020; Hansen e al.
2015). Cu en UGI, howe e , may no be enough o cope wi h clima e
change (O’Neill e al. 2022).
En i onmen al jus ice esea ch sugges s pas unjus u ban de elop-
men p ac ices a e associa ed wi h p esen spa ially embedded dispa -
i ies o UGI in ci ies (Schell e al. 2020). S udies in New Yo k, San
An onio, and Chicago ound communi ies o colo and/o low income
communi ies ha e less access o UGI, such as ee canopy (Nyelele &
K oll 2020), and associa ed ecosys em se ices, such as lood and hea
mi iga ion (Yi e al., 2019; Gonz´
alez e al., 2022; He e os-Can is &
McPhea son 2021). The e is a need o inc ease equi able dis ibu ion,
quan i y, and quali y o ecosys em se ices o cu en and u u e gen-
e a ions (Yi e al. 2017; Nyelele e al. 2019).
UGI’s capaci y o locally supply ecosys em se ices makes i a ele-
an s a egy o ad ance en i onmen al jus ice, p e en u u e ha m,
and lessen se e i y o clima e haza ds (Hoo e e al. 2021). Ye , despi e
eco ded en i onmen al injus ices, municipal planning and p io i iza-
ion app oaches o UGI do no sys ema ically conside i s spa ially-
explici , u u e jus ice implica ions ollowing implemen a ion, o en
a ibu ed o lack o adequa e go e nmen al mechanisms (Chu &Can-
non, 2021; Hoo e e al. 2021; Hoo e e al. 2023; G abowski e al.
2023). In he nea - e m, municipal clima e- ela ed in es men s a e
p io i ized among a b oade scope o o he capi al in es men s, whe e a
mul i ude o ac o s a he p esen ime mus coalesce o a p ojec o be
unded, designed, and implemen ed. In his con ex , jus ice-o ien ed
UGI planning o en ocuses on whom he p ojec would bene i i
implemen ed oday, a he han h ough i s asse li e. Ye , unde s and-
ing he u u e spa iali ies o UGI implemen a ion is impo an so as o
no magni y exis ing inequi ies o p oduce new injus ices. Unde s and-
ing u u e UGI can help p eemp i ely plan o mi iga e g een gen i i-
ca ion, whe e implemen a ion absen o o he in e en ions can push
ou communi ies due o ising cos s and changing neighbo hood cha -
ac e (Jimenez &Maan ay 2023). Gi en he inequi able dis ibu ion o
UGI om pas decisions (Schell e al. 2020), we mus comp ehensi ely
plan and implemen UGI o p o ide ecosys em se ices o lessen cu en
and u u e injus ices (Kabisch and Haase 2014; Rigolon e al. 2018).
2. Resea ch objec i es
Ou esea ch aims o unde s and how he an icipa o y planning o a
mul i unc ional land use s a egy—UGI—can con ibu e o alle ia ing
en i onmen al injus ices and mi iga ing clima e haza ds in ci ies. We
ask: wha a e he po en ial en i onmen al jus ice implica ions o
ecosys em se ices p o isioned om UGI in a co-p oduced u u e sce-
na io compa ed o a business-as-usual u u e scena io? We answe his
by: (1) assessing capaci y o UGI o p o ision lood and hea mi iga ion
ac oss cu en (2016), business-as-usual, and co-p oduced o clima e
esilience u u e (2100) scena ios using spa ially-explici models; (2)
assessing he change o ecosys em se ice p o isioning om UGI be-
ween cu en and al e na i e u u e scena ios; and (3) assessing he
po en ial en i onmen al jus ice adeo s o ecosys em se ices ac oss
cu en and al e na i e u u e scena ios.
3. Me hods
3.1. Co-p oducing u u e scena ios in New Yo k ci y
We use New Yo k Ci y (NYC) as a case s udy. NYC is he mos
populous ci y in he Uni ed S a es, wi h o e 8.8 million esiden s in a
M. Du a e al. Landscape and U ban Planning 254 (2025) 105249
2
o al a ea o 783.73 km
2
making up i e bo oughs (Manha an, Queens,
B onx, B ooklyn, and S a en Island). Clima e haza ds pose a isk o NYC,
including looding and ex eme hea (NYCEM 2019). The New Yo k
Panel on Clima e Change (NPCC) high ange p ojec ions sugges mean
annual empe a u es could ise by mo e han 3.5 ◦C and mean annual
p ecipi a ion could inc ease by 17% by he 2050s (B aneon e al. 2024;
Gonz´
alez e al. 2019).
To en ision a u u e NYC esilien o clima e haza ds, we collabo-
a ed wi h he NYC Mayo ’s O ice o Clima e and En i onmen al Jus ice
o lead he NYC Clima e Adap a ion wo kshop se ies. Local, s a e, and
ede al NYC go e nmen s akeholde s om 24 agencies (Supplemen a y
Ma e ial (SM) Table 1) a ended i e 2.5-hou i ual wo kshops in Fall
2021 consis ing o an icipa o y planning ac i i ies o add ess clima e
esilience. We in i ed go e nmen ep esen a i es speci ically wi h he
goal o b idge silos and os e c oss-agency clima e esilience and
adap a ion discussion and planning. Th ough hese wo kshops, pa ici-
pan s co-de eloped six u u e scena ios o NYC 2100 o add ess clima e
haza ds, including one add essing ex eme hea . The goal o scena io
building was o en ision posi i e and esilien u u es, di e en om
ypical business-as-usual app oaches, by adically ans o ming social,
ecological, and buil in as uc u e o a long- e m, u u e NYC. The yea
2100 was chosen o c ea e an uncommon space o mo e ans o ma i e,
long- e m isioning and goal se ing ha isn’ cons ained by cu en
ci cums ances. Following me hods in Iwaniec e al. (2020a) and Cook
e al. (2022), in he i s wo kshop, pa icipan s wo ked in small g oups
(app oxima ely six people ep esen ing di e se agencies/sec o s, plus a
ained acili a o who guided he ac i i ies) o de ine long- e m goals
o hei scena io in 2100 (e.g., elimina e hea - ela ed illness and mo -
ali y by 2100). In he ollowing wo kshop, pa icipan s con inued o
wo k in he same small g oups and used backcas ing app oaches o co-
de elop implemen a ion s a egies (e.g., expand UGI h oughou
ulne able neighbo hoods) needed o achie e hei scena io’s goals (Box
1). The emaining wo kshops ocused on ac i i ies o e ine he de ails
on spa ial loca ions, imelines, and go e nance app oaches o achie e
he goals (Table 1) as well as engage collec i e eedback om o he
g oups. In his pape , we assess he ex eme hea scena io as hea is he
deadlies clima e- ela ed haza d in New Yo k Ci y (Ma e e al. 2024),
and UGI has measu able bene i s o cooling in ci ies.
Table 1
Co-p oduced goals and s a egies o add ess hea esilience en isioned by pa icipan s in co-p oduc ion wo kshop (Fall 2021). Each s a egy ha add essed spa ially
explici land co e was ansla ed in o an addi ional modeling ule o he cellula au oma a (CA) land use land co e (LULC) model. Modeling ules desc ibe he
ansi ion ules o e e y 4x4m cell om one land co e (e.g., mixed use u ban) o ano he (e.g., open space).
Co-p oduced hea esilience goals o 2100 S a egies o add ess goals as desc ibed by pa icipan s in
isioning wo kshop
T ansla ion o spa ially explici s a egies in o
LULC modeling ules
(1) Elimina e hea ela ed illness and mo ali y, and
(2) Minimize and euse was e hea sou ces and a
same ime maximize ci y g een and blue
in as uc u e ( ege a ion and wa e ea u es)
Inc ease g een co ido s along oadways
−50% o s ee pa king spaces a e epu posed by 2025 o g een
space/s ee ees
−No p i a e s ee pa king by 2035
−Reclaim cu b space o g een in as uc u e (less pa king,
ewe ca s) by 2035
1. Na ow s ee s by adding g een co ido . Allow
ansi ion o u ban g een/ o es s co ido s along s ee s
by 2030; allow 8 m o g een co ido s on s ee s <4
lanes and 4 m co ido s on 2 lane s ee s
Inc ease wa e ea u es o e apo a i e cooling (mis e s o
ponds)
−15% o neighbo hood space includes wa e ea u es
2. Allow ansi ion om ege a ion classes o wa e in
medium and low densi y esiden ial (minimum a ea se
o 5x5m (25 m
2
))
Ensu e ha all New Yo ke s a e well-se ed by g een and
e icien public ansi
−Expanded access o people-based, be e ne wo ked public
ansi
−Ca ee zones ( inancial dis ic ), 80% educ ion in uck
a ic
−Bike in as uc u e p io i ized o e p i a e ca s wi h >50%
a el lanes o bikes
−Mo e subways, Bus Rapid T ansi , &be e ne wo k
3. Ca ee zones c ea ed by ansi ion om s ee s/
impe ious o g een space (pa k) −- p io i ize majo
in e sec ions >4 lanes nea majo ansi hubs; include
in e sec ions in all bo oughs
Inc ease g een in as uc u e h oughou ci y, especially
hea ulne able a eas
−By 2025, unding o s ee ees, g een oo s, &g een
in as uc u e in ulne able a eas
−By 2030, 2% ci y budge o g een space
−By 2035, new s ee ee bed designed o s o mwa e cap u e
−By 2040, eclaim cu b space o g een in as uc u e (less
pa king, ewe ca s)
−By 2050, 50% inc ease in s ee ees
−By 2060, 90% o oo space is g een o sola
−By 2065, all s ee ee beds include s o mwa e cap u e
−Inc ease g een space &na u al in as uc u e (including g een
acades) by 50%
4. In addi ion o ule 1 (abo e), inc ease ansi ion o
g een space om u ban/ esiden ial, inc ease o es
pa ches along oadways/s ee co ido s o ep esen
50% inc ease o s ee ee, inc ease ansi ions o
u ban g een oo s (using s a is ical eg ession based on
lo size &income o whe e g een oo s a e likely o
happen)
Inc eased shade / canopy co e h oughou ci y
−Upda e codes o include shade equi emen s in new
cons uc ion, plazas, open space
−50% o open space is shaded
5. In addi ion o ule 4 (abo e), ansi ion om open
space o o es ed o shade.
Ta ge neighbo hoods based on NYC Hea Vulne abili y
Index (HVI) and map*
6. S a wi h HVI neighbo hoods o g een space
ansi ions. In 2050, open ansi ions o ull ci y.
−apply his o ule o inc ease in s ee ees
−inc ease in wa e ea u es
−inc ease in ansi ion om u ban o g een space
* HVI da a publicly a ailable: h ps://a816-dohbesp.nyc.go /Indica o Public/da a- ea u es/h i/.
M. Du a e al. Landscape and U ban Planning 254 (2025) 105249
3
3.2. Fu u e land use land co e simula ions
In o de o explo e ou comes o he en isioned u u e, p ojec ions o
land use and land co e (LULC) we e simula ed o he co-p oduced
2100 hea scena io and a 2100 business-as-usual scena io. Following
me hods om Mus a a and colleagues (2018, 2021), he p ojec ions a e
simula ed om a mul i-objec i e Ma ko Chain Mon e Ca lo (MO-MCM)
cellula au oma a (CA) model, which di ides he spa ial a ea in o a g id
o cells and each cell’s LULC e ol es o e disc e e annual ime s eps
based on p ede ined ules and he s a es o neighbo ing cells. The base
CA model was ained wi h da a om 2002 and 2016 MapPLUTO land
use classi ica ions a a 4-m spa ial esolu ion o es ima e he a e o
change om one land use s a e o ano he a each imes ep. The model
also accoun s o popula ion and geophysical a iables, such as ele a-
ion, slope, and oad ne wo ks, ha we e gene a ed a o esampled o 4-
m esolu ion o align wi h MapPLUTO da a (SM Table 2).
Building on his base CA model, he models o business-as-usual
(BAU) and he co-p oduced scena ios p ojec u u e LULC change. The
BAU scena io was modeled as a baseline o poin o compa ison o he
co-p oduced scena io in 2100 (simila o Ahmadisha a e al. 2020;
Bollige e al. 2008). The BAU model ep esen s a con inua ion o
ex apola ion o ecen (2002–2016) LULC ends—p ojec ed o
2100—based on he quan i y o changes om one land use class o
ano he and he spa ial loca ions o obse ed changes (Fig. 1, panel 1).
The co-p oduced scena io LULC p ojec ions also build upon he base CA
model, bu inco po a e addi ional modeling ules ha ep esen he
hea - ela ed spa ially explici s a egies de eloped du ing he wo kshop.
The scena io’s co-p oduced imelines we e used o de e mine he an-
si ion a es o espec i e changes and modeling ules, ollowing me hods
om O iz e al. (2021). The addi ional modeling ules accoun o
biophysical (e.g., 30% inc ease in ee co e along oadways by 2100)
and/o in as uc u e changes (e.g., epu pose 50% s ee pa king by
2030). (Fig. 1, panels 2–4; Table 1).
Bo h he BAU and co-p oduced models a e simply p ojec ing u u e
LULC changes. In ou case, he model does no aim o p edic how NYC
will look in he u u e, as i does no cap u e u u e socio-poli ical dy-
namics, such as gen i ica ion, no inco po a e demog aphic p ojec ions
o po en ial clima e change in luences o he LULC, such as sea le el ise.
Howe e , he LULC models allow us o explo e he consequences o
po en ial decisions on he dis ibu ion o u u e ecosys em se ices
(desc ibed in sec ion 2.2). The ou pu o scena io modeling consis ed o
LULC maps o each scena io, consis ing o 14 ypes o he 2100 BAU
and 18 ypes o he 2100 hea scena io, whe e an addi ional ou ypes
we e included o di e en ia e buildings wi h g een oo s (SM Fig. 1).
3.3. Ecosys em se ice assessmen
We assessed he capaci y o UGI in NYC o supply wo ecosys em
se ices: lood and hea mi iga ion. To quan i y his, we used he In e-
g a ed Valua ion o Ecosys em Se ices 3.12.0 (InVEST) oolse (Na u al
Capi al P ojec 2022), a spa ially-explici open-sou ce modeling so -
wa e ha p o ides biophysical es ima ions o ecosys em se ice supply.
InVEST o u ban ecosys em se ices ha e been ho oughly p esen ed in
o he s udies such as Hamel e al., 2021. He e, we b ie ly desc ibe i s use
and unc ioning and p o ide ex ended ma e ials in SM 1.1 and 1.2.
3.3.1. Flood mi iga ion
The InVEST U ban Flood Risk Mi iga ion model calcula es he ca-
paci y o UGI o e ain uno du ing a speci ic s o m e en (Na u al
Capi al P ojec 2022). Based on land use cha ac e is ics ac oss NYC
including uno po en ial (SM Table 3; USDA 2004) and soil ype, we
calcula ed he m
3
o uno e ained du ing a 24-hou , 5-inch (127 mm)
ain e en , as an example o a p esen ex eme (app oxima ely a 100-
yea e en ) s o m (SM Sec ion 1.1;Gonz´
alez e al. 2019). Each sce-
na io’s as e ou pu (4-m pixels ep esen ing uno e en ion olume in
m
3
) was agg ega ed by a e aging alues wi hin he census block g oups
(CBG) o compa e uno e en ion and sociodemog aphic a iables.
3.3.2. Hea mi iga ion
The InVEST U ban Cooling model calcula es he capaci y o u ban
g een in as uc u e o cool he su ounding a ea. Fo his model, land
use class cha ac e is ics on shade, e apo anspi a ion, albedo, and dis-
ance om la ge g eenspaces (SM Table 4) we e used o model he hea
mi iga ion index (HMI). HMI is a uni less measu e ha ep esen s
cooling capaci y and ep esen s he p opo ional capaci y o a pixel o
cool ela i e o he maximum u ban hea island e ec obse ed in NYC
(SM Sec ion 1.2;Na u al Capi al P ojec 2022). The ou pu as e s (4-m
pixels ep esen ing HMI) we e agg ega ed and a e aged o he CBG scale
o compa e he HMI and sociodemog aphic a iables.
3.4. En i onmen al jus ice indica o s
To e alua e en i onmen al jus ice and ecosys em se ices in New
Yo k Ci y, we agg ega ed 2020 sociodemog aphic a iables (SM
Table 5) a he CBG scale, he smalles sha ed spa ial scale among
sociodemog aphic a iables. We ecognize demog aphic dis ibu ions
may change by 2100, howe e mos popula ion p ojec ions a e sho -
e m, do no always inco po a e sociodemog aphics, and do no exis
a a ine enough spa ial scale o meaning ul analysis (Balk e al. 2022).
Thus, u u e popula ion p ojec ions add ano he laye o unce ain y o
ou analysis. Ins ead, we e alua e how u u e land co e and UGI
changes by 2100 may impac cu en communi y demog aphics.
Fu he mo e, municipal UGI planning la gely conside s p esen pop-
ula ions o implemen a ion jus i ica ion— a he han u u e pop-
ula ions—allowing us o p o ide ealis ic me hodology.
New Yo k Ci y has 6391 inhabi ed CBGs; we compiled in o ma ion
om bo h he decennial census and Ame ican Communi y Su ey (ACS)
es ima es (He e os-Can is and McPhea son, 2021). Wi h he 2020
census, we calcula ed he p opo ion o indi iduals li ing in a CBG
iden i ying as whi e, Black and/o A ican-Ame ican, Hispanic and/o
La ino, and Asian ( he p edominan acial/e hnic ca ego ies) using he
Tidycensus package in R (4.1.2). Using he 2016–2020 ACS 5-yea es-
ima es (U.S. Census Bu eau n.d.) and Tidycensus package (Walke &
He man 2023) in R (4.1.2), we compiled and calcula ed he p opo ion
Box 1
Wo kshop pa icipan s co-de eloped a ision o NYC in 2100 ha is esilien o ising empe a u es and ex eme hea . The u u e scena io
add esses wo main goals o (1) elimina ing hea - ela ed illness and mo ali y, while (2) minimizing and eusing was e hea sou ces and
maximizing UGI. Hea - ela ed illness is educed h ough mic o-cooling cen e s, egula ions o ou doo wo ke s and indoo empe a u es, and
inc eased communi y heal h p og ams. The scena io maximizes he UGI and wa e ea u es, including 50% shaded open space and 50% mo e
g een co ido s by 2100, eclaimed cu b space o ee canopy, 25% o neighbo hood space includes wa e ea u es, and 90% o e o i and new
oo space is g een o sola by 2060. The scena io minimizes new ene gy use by educing o eusing hea was e (e.g., om ai condi ioning) due
o upda ed building codes, passi e o low-ene gy building design and ma e ials, and equi ed g een oo e o i s (mo e de ails and imelines can
be ound in Cook e al. 2022).
M. Du a e al. Landscape and U ban Planning 254 (2025) 105249
4
o indi iduals pe CBG o he emaining sociodemog aphic a iables:
pe cen in po e y, emale-headed households, uninsu ed heal h
co e age, no in e ne access, low educa ional a ainmen , and en e s
(SM Table 5). All da a we e inco po a ed in o he geog aphic bounda ies
o NYC CBGs ob ained om he TIGER/LINE da abase ia he U.S.
Census Bu eau (n.d.).
3.5. En i onmen al jus ice implica ions o ecosys em se ices
We e alua ed he en i onmen al jus ice implica ions o lood and
hea mi iga ion ac oss scena ios by assessing how sociodemog aphic
a iables p edic ecosys em se ice dis ibu ion in NYC h ough s ep-
wise eg ession modeling. O dina y leas squa es (OLS) and geog aph-
ically weigh ed eg ession (GWR) models can be employed as an
explo a o y mechanism o examine how pa icula sociodemog aphic
indica o s co ela e wi h en i onmen al bene i s o bu dens (Gilbe &
Chak abo y 2011; Schwa z e al. 2015; Yu e al. 2023). The OLS
eg ession models ep esen global ela ionships o en i onmen al jus-
ice ac oss a da ase , in ou case highligh ing he subse o sociodemo-
g aphic a iables impo an in p edic ing ecosys em se ices ega dless
o spa ial dis ibu ion. The second s ep GWR p o ides addi ional un-
de s anding o how ci ywide ela ionships, as es ablished in he OLS,
may a y ac oss space.
Fo he s epwise modeling, we i s conduc ed a Spea man’s co e-
la ion es o explo e ela ionships be ween he p edic o sociodemo-
g aphic a iables (Yi e al., 2019). I wo sociodemog aphic p edic o s
showed a ela ionship o 0.5 o g ea e , he wo a iables we e no
ini ially included in he same eg ession model (Shaw e al. 2023). We
hen employed o dina y leas squa es (OLS) eg ession models, using a
o wa d and backwa d s epwise app oach and selec ing he model wi h
he lowes Akaike In o ma ion C i e ion (AIC). We added sociodemo-
g aphic a iables p e iously emo ed due o co ela ions equal o o
g ea e han 0.5 and kep hem in he model i he AIC o he model
dec eased and he a iance in la ion ac o s (VIF) did no exceed 5
(Snee, 1973). OLS a e based on an assump ion o s a iona i y and
homogenei y–o , ha he global pa ame e s beha e iden ically ac oss
he s udy a ea (Mennis 2006). The OLS bes i models se ed o iden i y
he combina ion o sociodemog aphic a iables ha bes explained he
ange o ecosys em se ice alues ac oss NYC. Gi en he co-p oduc ion
p ocess ocused on making NYC mo e hea esilien as a single
geog aphic egion, his analysis p o ides us in o ma ion on he ac o s
ha impac ci ywide jus ice and ecosys em se ice ou comes.
To accoun o he spa ial a iabili y wi hin he s udy egion, we
ollowed he OLS wi h a GWR. We conduc ed a local Mo an’sI es on
he OLS models and con i med spa ial au oco ela ion ac oss he sce-
na ios—indica ing he global OLS models and hei associa ed en i-
onmen al jus ice p edic o s may a y ac oss NYC. The e o e, we also
employed GWR using he spgw package (Bi and &Yu, 2022) in R
(4.1.2) o explo e local a iabili y wi hin NYC. The OLS eg ession
equa ions a e he basis o his p ocess whe e GWR analyzes a gi en
eg ession equa ion sepa a ely ac oss op imal numbe s o CBGs wi hin
NYC, de ined using an adap i e ke nel (Gilbe &Chak abo y 2011).
3.6. Change in en i onmen al jus ice implica ions
We assessed adeo s be ween scena ios by compa ing he changes
in ecosys em se ice p o isioning be ween he p esen and u u e sce-
na ios and hei po en ial implica ions o en i onmen al jus ice. We
c ea ed “del a maps”showing he di e ence in cu en o u u e se ice
p o isioning capaci y (e.g., 2016 as e −2100 BAU as e o each
ecosys em se ice) and calcula ed he mean del a alue a he CBG le el.
Compa isons we e conduc ed be ween 2016 and BAU and 2016 and he
hea scena io. We hen conduc ed a Spea man’s co ela ion es o
explo e a baseline o associa ion wi h sociodemog aphic a iables,
culmina ing in a o wa d and backwa ds s epwise OLS eg ession using
he same AIC and VIF app oach as s a ed abo e. We ocused on ci ywide
ela ionships o his analysis o explo e he o e all ends in empo al
changes.
4. Resul s
4.1. Spa ial dis ibu ion o ecosys em se ices
4.1.1. Flood mi iga ion
Ou esul s show spa ial a ia ions in he capaci y o UGI o mi iga e
looding du ing an ex eme s o m (5-in. o e a 24-hou pe iod, Figs. 2 &
3) ac oss scena ios. While he mean uno e en ion in CBGs ac oss NYC
does no a y d as ically be ween 2016 NYC and 2100 BAU NYC (0.33-
Fig. 1. Schema ic o de eloping he u u e scena io LULC ou pu s and InVEST analysis o u u e ecosys em se ices. The mul i-objec i e Ma ko Chain Mon e Ca lo
(MO-MCM) cellula au oma a (CA) LULC model was ained ( ollowing Mus a a e al 2018) wi h 4-m esolu ion LULC da a, along wi h o he baseline da a, o p ojec
he 2100 Business-as-Usual LULC ou pu . Addi ional modeling ules we e added based on he co-p oduced s a egies o add ess hea esilience in o de o p ojec he
2100 Hea Resilience LULC ou pu . The LULC ou pu s eed in o U ban InVEST ecosys em se ice models. Addi ional de ails on LULC classes in SM Fig. 1 and how
wo kshop s a egies we e ansla ed in o modeling ules in Table 1.
M. Du a e al. Landscape and U ban Planning 254 (2025) 105249
5
m
3
o 0.34-m
3
), he 2100 co-p oduced hea scena io mean uno
e en ion inc eases by 55%, o 0.58-m
3
, compa ed o 2016 Table 2). The
majo i y o CBGs do no d as ically change posi i ely o nega i ely in
uno e en ion be ween 2016 and 2100 BAU NYC (mean =0.00+/-
0.07-m
3
;Table 2), whe eas he majo i y o CBGs inc ease (mean =
0.25+/-0.1-m
3
;Table 2) in uno e en ion capaci y om UGI be ween
2016 and 2100 co-p oduced hea scena io.
4.1.2. Hea mi iga ion
Ou esul s show he capaci y o UGI o mi iga e hea in NYC a ies
spa ially (Fig. 2) and be ween scena ios (Fig. 4). Whe e hea mi iga ion
index is a uni less measu e o cooling om ege a ion on a scale om
0 o 1, and ep esen s he p opo ional capaci y o a CBG o cool ela i e
o he maximum u ban hea island e ec obse ed in NYC, he spa ial
dis ibu ion o highes hea mi iga ion in NYC occu s in and su ounding
la ge g een spaces (such as Cen al Pa k in Manha an and Fo es Pa k in
Queens). While he mean is 0.11 o bo h 2016 NYC and 2100 BAU NYC,
he mean is 108% highe a 0.37 o he 2100 co-p oduced hea scena io
(Table 2). In his co-p oduced scena io, HMI alues a e mo e e enly
dis ibu ed compa ed o ha o he o he scena ios.
Compa ing 2016 and 2100 BAU NYC, he majo i y o HMI alues do
no change ac oss CBGs (Fig. 4). The ange o changes ac oss CBGs a e a
a g ea e magni ude gained han los (Table 2). On he o he hand, he
majo i y o CBGs inc ease in HMI alues be ween 2016 NYC and he
2100 co-p oduced scena ios (mean =0.27+/-0.06; Table 2). A hand ul
o CBGs on he bounda ies o NYC, such as in Jamaica Bay and S a en
Island, appea o make up he majo i y loss o hea mi iga ion capaci y
(Fig. 4).
4.2. Reg essions o ecosys em se ice spa ial associa ions
Ou esul s illus a e unique associa ions be ween en i onmen al
jus ice and he dis ibu ion o ecosys em se ices in each scena io as well
as change in p o isioning be ween cu en and u u e scena ios. In o al,
ou analyses esul ed in 10 bes i ela ionships, as iden i ied in he
s epwise app oach ( i e pe ecosys em se ice, Table 3). Each bes i
model had a di e en combina ion o en i onmen al jus ice indica o s
(SM Sec ion 2), whe e collec i ely all a iables included in his s udy
we e signi ican in a leas a subse o he models (Table 3). The mos
common indica o s in he models included pe cen Asians, which p e-
dic ed se ice p o ision in all he models, and pe cen en e s and
popula ion densi y which we e signi ican in all bu wo models
(Table 3). Po e y, emale-headed household, and no high school
diploma we e he leas common a iables in he bes i models.
Globally, he posi i e o nega i e ela ionship and magni ude o
coe icien s a ied ac oss models (Table 3). The 2016 and 2100 BAU
hea mi iga ion OLS models con ained he la ges posi i e coe icien s
o he a iables whi e, Hispanic/La ino, and Black/A ican-Ame ican
(SM Table 6). O e all, sociodemog aphic a iables ela ed o acial/
e hnici y we e mos commonly p edic o s ac oss scena ios. No ably, he
adjus ed R
2
alues o all OLS models a e ela i ely small (Table 3, SM
Table 7), indica ing he en i onmen al jus ice a iables p edic only up
o 10% o a ia ion in he ecosys em se ice models a he ci ywide
scale. Ou analyses illus a e ela ionships be ween en i onmen al
Fig. 2. Dis ibu ion o uno e en ion (m
3
, op panel) and hea mi iga ion (uni less, bo om panel) p o isioned by UGI ac oss h ee NYC scena ios (NYC in 2016,
BAU in 2100, and co-p oduced scena io o add ess hea in 2100). Da a p esen ed as census block g oup’s a e age uno e en ion o hea mi iga ion capaci y pe 4-
m
2
pixel.
M. Du a e al. Landscape and U ban Planning 254 (2025) 105249
6
jus ice and ecosys em se ices, bu indica e sociodemog aphics a e no
he main d i e s o p o isioning a he ci ywide scale.
To be e unde s and how he global eg ession models a ied ac oss
space, we used geog aphically weigh ed eg essions (GWR) o explo e
he spa ial he e ogenei y o ela ionships be ween ecosys em se ices
and sociodemog aphic a iables ac oss NYC. The GWR models o lood
and hea mi iga ion did no ully accoun o spa ial clus e ing in he
models, howe e , hey did educe he spa ial dependence in he e-
siduals om he OLS models in mos cases and imp o ed pe o mance
(global R
2
) ac oss models (Table 4). Speci ically, he GWR models
de ec ed locally a ying associa ions ac oss space gi en he global
eg ession models iden i ied h ough he OLS app oach. Ou esul s
show a ange o spa ially s ong and weak associa ions (local R
2
) o
bo h ecosys em se ices ac oss CBGs pe scena io (Fig. 5, SM Table 8).
O e all, he GWR models mos s ongly p edic lood (adjus ed R
2
=
0.57) and hea (adjus ed R
2
=0.92) mi iga ion in he 2100 co-p oduced
scena ios, compa ed o hei espec i e 2016 and BAU scena ios
(Table 4). Gi en s epwise p ocess, simila en i onmen al jus ice p e-
dic o s a e impo an in he GWR models as he OLS models (SM
Table 8). Howe e , he GWR esul s highligh spa ial a ia ion, whe e
o example he s onges ela ionships (highes R
2
) o lood mi iga ion
and sociodemog aphic a iables occu in coas al a eas o sou h B ooklyn
and Queens and no he n Manha an and he B onx (Fig. 5). The s ong
ela ionships be ween hea mi iga ion and sociodemog aphic a iables
a e mo e e enly dis ibu ed ac oss NYC (Fig. 5). Ac oss GWR models, he
ange o coe icien s (SM Table 8) is g ea es o he ace/e hnici y
a iables, compa ed o o he sociodemog aphic a iables, such as no
educa ion, en e , and heal h insu ance s a us. This coe icien ange
highligh s how ace/e hnici y g oups may speci ically expe ience high
a ia ion in access o UGI h oughou he ci y in a gi en scena io. Fo
example, while some Asian communi ies would be associa ed wi h
highe access o UGI om he p oposed s a egies in he co-p oduced
scena ios, he e will be o he loca ions whe e Asian communi ies will
expe ience lesse access o UGI gi en land use changes.
5. Discussion
The inc ease in equency and se e i y o clima e-d i en haza ds in
ci ies necessi a es he s a egic planning and implemen a ion o solu-
ions ha bene i he communi ies and loca ions mos a isk (Chu &
Cannon 2021). U ban clima e esilience e o s, howe e , a ely ake a
long- e m planning app oach o de elop sha ed, posi i e isions o hen
explo e po en ial adeo s ahead o implemen a ion (Cook e al. 2022;
Iwaniec e al., 2020a; Mee ow 2020). Ou esea ch seeks o ill his gap
by assessing how he an icipa o y planning o UGI can con ibu e o bo h
he alle ia ion o en i onmen al injus ices and clima e haza d
Fig. 3. Change in mean uno e en ion capaci y ac oss census block g oups be ween 2016 NYC and 2100 BAU NYC ( op panel) and be ween 2016 NYC and 2100 co-
p oduced hea scena io (bo om panel). Nega i e alues (o ange) indica e a lowe capaci y o ecosys em se ices in he u u e, and posi i e alues (pu ple) indica e
highe u u e capaci y. Zoomed in a eas (clockwise) highligh key NYC neighbo hoods wi h se ice change and a ying p edominan demog aphics, including 1:
p edominan ly Hispanic/La ino and Black/A ican-Ame ican; 2: p edominan ly Asian; 3: p edominan ly Black/A ican-Ame ican; 4: p edominan ly whi e. (Fo
in e p e a ion o he e e ences o colo in his igu e legend, he eade is e e ed o he web e sion o his a icle.)
M. Du a e al. Landscape and U ban Planning 254 (2025) 105249
7
mi iga ion in NYC. Ou esul s highligh , howe e , no all communi ies
may bene i equally gi en planned s a egies, indica ing need o addi-
ional planning in e en ions o add ess emaining en i onmen al jus-
ice challenges.
Simila o o he esea ch assessing he p esen inequi able dis ibu-
ion o ecosys em se ices in NYC (He e os-Can is &McPhea son,
2021; Nyelele &K oll, 2020; Nyelele, K oll, &Nowak, 2019), we ound
p esen day (2016) ecosys em se ice p o isioning in NYC is inequi ably
dis ibu ed. Communi ies o colo and low-income communi ies a e,
o e all, associa ed wi h lowe access o ecosys em se ices p o isioned
by UGI, whe eas weal hie and/o whi e communi ies ha e disp o-
po iona ely high access. Gi en p esen day injus ices and a p edic ed
inc eased equency and se e i y o clima e haza ds (Gonz´
alez e al.
2019; Ma e e al. 2024; Solecki &Rosenzweig 2019), i is c i ical o
unde s and how u u e de elopmen and an icipa o y planning may
alle ia e o ein o ce hese quali ies.
Ou wo old ci ywide and spa ially-explici eg ession analyses p o-
ide no el in o ma ion on ecosys em se ice access and dis ibu ion
ac oss communi ies and among ou comes o po en ial u u e planning
decisions. The esul s p o ides c i ical in o ma ion needed o iden i y
adeo s and a oid unwan ed implica ion h ough comp ehensi e
analysis p io o implemen a ion. A a ci ywide scale, ou esul s high-
ligh how o en acial/e hnic ca ego ies and en e s a us (a p oxy o
po e y) ma e mos o p edic ing se ice p o ision and changes in
access oday and in he u u e, depending on wha ac ion is aken. This is
c i ical o planne s and policymake s because decisions a e ypically
in o med gi en ci ywide ends obse ed. Ye , ou spa ially-explici
analyses highligh he impo ance o explo ing how ci ywide ends
a y among neighbo hoods o a oid a na a i e ha pe pe ua es a
monoli hic expe ience o en i onmen al injus ice. Fo decision-make s
iden i ying whe e o implemen change unde minimal esou ces, ou
esul s help answe whe e and o whom his can be done o ensu e he
bigges bene i s a e ealized bo h ci ywide and locally.
Mo e speci ically, a he ci ywide scale compa ing p esen (2016)
and he u u e BAU scena io in 2100, ou esul s e eal he a e age
change o bo h hea and lood mi iga ion is ze o. This sugges s while
UGI and associa ed se ices inc eased in some loca ions o he NYC BAU
scena io, hose se ices simila ly dec eased elsewhe e, leading o no
o e all ci ywide cumula i e change in se ice p o isioning. Impo -
an ly, ou esul s illus a e ce ain sociodemog aphic g oups a e asso-
cia ed wi h less and diminishing access o p o isioning gi en BAU. Ou
ci ywide models on change in UGI om 2016 o BAU e eals Black/
A ican Ame ican communi ies expe ience a loss in lood mi iga ion
while Asian communi ies a e associa ed wi h a loss in bo h lood and
hea mi iga ion. We addi ionally see a loss in se ices o hose wi hou
in e ne and heal h insu ance. Ou esul s explo ing change be ween
2016 and BAU sugges such a de elopmen ajec o y would no alle-
ia e injus ices and may exace ba e o econ igu e he une en dis i-
bu ion o p o isioning o some g oups. No ably, he equency,
in ensi y, and cycles o ecu ing ex eme haza ds in NYC is expec ed o
inc ease and sys ema ically inc ease he o e all ulne abili y o NYC
communi ies o e ime (Solecki &Rosenzweig 2019; Gallina e al.
2016). In his sense, ou esul s may unde es ima e he se e i y o u u e
haza ds and o e emphasize he capaci y o UGI o p o ide ecosys em
se ices du ing u u e ex eme e en s.
Howe e , compa ing p esen (2016) and he co-p oduced, u u e
scena io in 2100 a he ci ywide scale illus a es he p omise o in en-
ional, an icipa o y planning o u u e clima e esilience. The co-
p oduced scena io e eals an a e age inc ease in bo h lood and hea
mi iga ion in NYC. In e es ingly, hea mi iga ion is g ea e han lood
mi iga ion om UGI, sugges ing a iabili y o UGI o p o ide clima e
bene i s. O e all, his wo k highligh s how c oss-sec o al, an icipa o y
planning can a ge a speci ic challenge in o de o gain bene i s o
clima e esilience o e all and especially ad ances o ex eme hea
mi iga ion goals. Ye , s ill, hose changes in clima e esilience be ween
p esen day and he co-p oduced u u e ha e di e gen implica ions o
communi ies. Fo example, Asian communi ies and hose wi hou
in e ne o high school would lose access o lood se ices on a e age
ci ywide while emale-headed households, en e s, hose wi hou heal h
insu ance, and p edominan ly whi e communi ies would gain access. On
he o he hand, o hea mi iga ion, we ound a disp opo iona e in-
c ease in access o en e s (p oxy o po e y), p edominan ly Asian and
Black/A ican-Ame ican communi ies, and hose wi hou in e ne .
O e all, he ci ywide analyses indica e no all communi ies will bene i
equally and he mul idimensional na u e o ulne abili y. O e lapping
sys ems o opp ession and inequali y, including limi ed economic, po-
li ical, land, and social powe , o en con ibu e o clima e ulne abili y
(Malin &Ryde 2018). While ou analyses do no ully cap u e in e -
sec ional iden i ies, he inclusion o a b oade sui e o en i onmen al
jus ice a iables adds nuance in explo ing how in e locking sys ems o
opp ession in NYC can ha m communi ies h ough spa ial access now
and in he u u e. Planne s and policymake s should conside hese di-
mensions in c oss-sec o al, an icipa o y planning, design, implemen a-
ion o u u e UGI o clima e esilience.
Ou spa ially-explici analyses e eal how ci ywide ela ionships as
explo ed abo e a y locally—p o iding c i ical da a o hose decision-
make s seeking whe e and o whom o implemen UGI. In 2016, s ong
in e se ela ionships be ween en i onmen al jus ice a iables and
ecosys em se ices may be indica i e o unde lying his o ic inequi ies
and p ocesses embedded in he u ban ab ic, such as edlining o
disin es men , s ill d i ing en i onmen al injus ices and p esen dis i-
bu ion o UGI (Hoo e &Lim 2021). These a e s a egic loca ions o
UGI in e en ions o lessen en i onmen al injus ices and in o m whe e
implemen a ion could be p io i ized oday. Those a eas wi h weake
associa ions be ween sociodemog aphic a iables and se ice p o ision
may be indica i e o al eady succeeding, ongoing jus ice-based UGI in-
e en ions, such as inc eased g eenspace wi hin NYC Housing Au-
ho i y p ope ies o low-income a eas. As we see in bo h he 2100 BAU
and co-p oduced scena ios, he s eng h o ela ionships be ween en i-
onmen al jus ice and ecosys em se ices a y o e space—highligh ing
he edis ibu ion o UGI may no ully disen angle pas injus ices. The
ou comes o he co-p oduced hea esilience scena io can be a good
s a ing poin o planne s o ee alua e whe e and when UGI
Table 2
Desc ip i e s a is ics o ecosys em se ices and en i onmen al jus ice indica o s
ac oss census block g oups. Fo he ows desc ibing change in ecosys em se ices
be ween 2016 and 2100 scena ios, nega i e alues indica e a educ ion o u u e
se ice capaci y, while posi i e indica es an inc ease in capaci y.
Min Max Mean (S d. de )
(1) Flood mi iga ion (m
3
)
NYC 2016 0.1 1.66 0.33 (0.1)
2100 NYC BAU 0.07 1.69 0.34 (0.09)
2100 NYC Hea 0.07 1.69 0.58 (0.13)
BAU −2016 −0.48 0.63 0 (0.07)
Co-p oduced hea scena io −2016 −0.42 0.86 0.25 (0.1)
(2) Hea mi iga ion index (uni less)
NYC 2016 0.03 0.68 0.11 (0.08)
2100 NYC BAU 0.04 0.7 0.11 (0.08)
2100 NYC Hea 0.09 0.74 0.37 (0.07)
BAU −2016 −0.19 0.5 0.01 (0.03)
Co-p oduced hea scena io −2016 −0.07 0.53 0.27 (0.06)
(3) En i onmen al jus ice indica o s
Asian p opo ion (%) 0 94 15 (18)
Black/A ican-Ame ican p opo ion (%) 0 91 20 (25)
Female-headed household (%) 0 100 35 (17)
Hispanic p opo ion (%) 0 96 28 (22)
No high school diploma, >age 25 (%) 0 42 3 (4)
No in e ne subsc ip ion a home (%) 0 100 13 (12)
Popula ion densi y (people pe CBG)* 4 8541 1311 (609)
Po e y (%) 0 100 13 (15)
Ren e (%) 0 100 63 (29)
Uninsu ed a e o heal h insu ance (%) 0 100 6 (7)
* CBGs wi h popula ion densi y ze o we e emo ed om analysis.
M. Du a e al. Landscape and U ban Planning 254 (2025) 105249
8
implemen a ion policies could occu , such ha u u e planning de-
cisions do no simply econ igu e injus ices.
P e ious s udies p ima ily explo e cu en o nea e m, bu no long-
e m u u e, en i onmen al jus ice implica ions o ecosys em se ices.
He e os-Can is and McPhea son (2021) iden i ied ecosys em se ice
supply–demand misma ch o cu en popula ions in NYC. Addi ional
s udies ha e linked legacies o esiden ial seg ega ion o cu en ineq-
ui able alloca ion o g een space (Rigolon e al. 2018) and ee canopy
(Nyelele &K oll 2020) o NYC communi ies o colo . Gonz´
alez and
colleagues (2022)—applying InVEST models o explo e cu en
ecosys em se ices in he Chicago egion—simila ly ound ha com-
muni ies o colo we e associa ed wi h ewe ecosys em se ices. Nea -
e m u u e s udies in A lan a and San An onio ound ha BAU sce-
na ios p ojec ed o 2050 also ein o ce exis ing inequi ies (Sun e al.,
2018; Yi e al., 2019). In con as , ou esea ch is no el as i builds on a
pa icipa o y planning p ocess and assesses he ways co-p oduced land
use s a egies o hea esilience may alle ia e en i onmen al injus ices,
which can ac as a ool o decision-make s o assess loca ions ha may
equi e an adjus men in UGI implemen a ion o a oid unwan ed
implica ions.
5.1. Planning conside a ions
NYC has ins i u ionalized and p io i ized clima e esiliency planning
and policy h ough mul iple a enues wi hin agencies. Pas ex eme
haza d e en s, such as Hu icane Sandy in 2012, Pos T opical Cyclone
Ida in 2021, and hea wa es in he summe o 2022, which led o he loss
o li es and p ope y in NYC, ha e a i med he salience o esilience
planning. PlaNYC, NYC’s clima e plan upda ed e e y ou yea s (NYC
MOCEJ 2023), e e ences UGI as a s a egy o inc ease clima e esil-
ience. Though NYC main ains s ong ambi ion o UGI—i s planning,
implemen a ion, o main enance is no consis en ly, o sys ema ically,
p io i ized alongside ci ywide no local a ia ions in en i onmen al
jus ice implica ions. Fo example, he NYC Pa ks Communi y Pa ks
Ini ia i e and MillionT eesNYC (NYC Pa ks 2023a; 2023b), had social
ulne abili y se ed as a ac o in implemen a ion. Ye , Ga ison (2021)
showed MillionT eesNYC did no su icien ly p io i ize unde se ed
communi ies o a measu able deg ee due o he legacies o maldis-
ibu ed g eenspace. Fu u e esea ch and s a egies mus b idge heo ies
o en i onmen al jus ice in o p ac ice by o ecas ing po en ial implica-
ions–bo h ci ywide and locally–ahead o implemen a ion o manage
adeo s, edi ec esou ces, o , a he e y leas , suppo anspa en
go e nance and communica ion (Ga ison 2021). Fo example, u u e
PlaNYCs can se he p ecedence o ac ionable en i onmen al jus ice by
con ening s akeholde s om mul iple sec o s in pa icipa o y se ings
and u ilizing amewo ks and pa icipa o y me hodologies, such as ou s,
o an icipa o y and spa ially s a egic planning wi h me ics o assess
po en ial ou comes p io o implemen a ion.
An icipa o y planning is c i ical o assess and limi unwan ed
adeo s. Se ings o con ene agencies ac oss a wide ange o
Fig. 4. Change in mean hea mi iga ion index capaci y ac oss CBGs be ween 2016 NYC and 2100 BAU NYC ( op panel) and be ween 2016 NYC and 2100 co-
p oduced hea scena io (bo om panel). Colo s and inse s as de ined in Fig. 3.
M. Du a e al. Landscape and U ban Planning 254 (2025) 105249
9
