National Institute for Environmental Studies (NIES)

Tomoko Hasegawa, Shinichiro Fujimori, Kiyoshi Takahashi and Toshihiko Masui

1

The 7th IAMC meeting @College Park, Nov 17-19, 2014

Background •  Climate  change  would  increase  the  number  of  malnourished  children.  

•  The  climate  impacts  strongly  depend  on  popula;on  and  GDP.  –  Based  on  the  SRES  and  CMIP3;  Need  updated.  –  Only  popula2on  and  GDP  were  considered;  other  socioeconomic  indicators  could  be  considered.    

•  A  new  interdisciplinary  scenario  framework  (SSP+RCP)  has  recently  been  designed  for  climate  change  research.  

•  Scenarios  for  various  fields  (e.g.  water  use)  have  been  developed  based  on  SSPs.  –  No  scenarios  for  risk  of  hunger  consistent  with  SSPs  have  been  developed.  

 2

Introduc2on

Objec;ves

1.  Develop  21st-­‐century  scenarios  for  the  risk  of  hunger  consistent  with  SSPs  as  a  basement  of  climate  impact  research  on  agriculture  

2.  Iden;fy  the  elements  strongly  affec;ng  future  risk  of  hunger  

•  7  socioeconomic  indicators  were  considered:  – Demographic  change,  GDP,  equality  of  food  distribu;on,    crop  yields,  irriga;on  area,  land  produc;vity  of  livestock  and  wood  products  

3

Introduc2on

Optimistic Pessimistic

Medium

Optimistic

Opt.(High income)

Med  (Middle income)

Pes.(low income)

SSP5Conventional  Development

SSP3Fragmentation

SSP2Middle  of  the  road

SSP1Sustainability

SSP4Inequity

Socioe

cono

mic  challenges  fo

r  mitigatio

n

Socioeconomic  challenges  for  adaptation

SSP challenges space Risk of hunger

Shared  Socioeconomic  Pathways

4

Low  popula;on  growth;  high  economic  growth;  

high  levels  of  educa;on  and  governance;  globaliza;on,  interna;onal  coopera;on,  technological  development,  

and  environmental  awareness.

Rapid  popula;on  growth;    Moderate  economic  growth;    low  levels  of  educa;on  and  

governance;    Regionaliza;on;  

 low  environmental  awareness.

 Low  popula;on  growth;  high  economic  growth;  

high  human  development;  low  environmental  awareness.

Method

Based  on  O’Neill  et  al.  (2014)

A  mixed  world,  with  rapid  technological  

development  in  high  income  countries.  In  other  regions,  development  proceeds  

slowly.  Inequality  remains  high.

•  Economic  model  •  Cover  the  en2re  economic  market  •  Fundamental  idea:    •  supply  =  demand  •  balanced  by  price  mechanism    Popula;on  &  income  growth  à  Increase  in  food  demand  à  Increase  in  food  price  à Producers:  increase  in  produc;on  

(cropland  expansion,  yield  growth)  à Consumers:  decrease  in  

consump;on;  shi_  to  less  expensive  goods  

AIM/CGE  (Computable  General  Equilibrium)  

Price

Quan;ty

Supply  curve Demand  curve

5

Domes;c  distribu;on  of  food  energy  (FAO,2008)

Method

Parameters  related  to  food  and  hunger

6

Method

AIM/CGE

Parameters: Popula;on,  demographic  change GDP Equality  of  food  distribu;on Crop  yield Irriga;on  area Land  produc;vity  of  livestock  and  wood  products Income  elas;city  of  food  demand Price  elas;city  of  land  use  change Price  elas;city  of  trade

Endogenous  variables: Food  consump;on Meat  consump;on Land  use  (Cropland,  pasture,  forest)

Optimistic Pessimistic

Medium

Optimistic

Opt.(High income)

Med  (Middle income)

Pes.(low income)

SSP5Conventional  Development

SSP3Fragmentation

SSP2Middle  of  the  road

SSP1Sustainability

SSP4Inequity

Socioe

cono

mic  challenges  fo

r  mitigatio

n

Socioeconomic  challenges  for  adaptation

SSP challenges space Risk of hunger

Each  parameters  were  determined  from  the  adapta2on  viewpoint

7

AIM/CGE

Parameters: Popula;on,  demographic  change GDP Equality  of  food  distribu;on Crop  yield Irriga;on  area Land  produc;vity  of  livestock  and  wood  products Income  elas;city  of  food  demand Price  elas;city  of  land  use  change Price  elas;city  of  trade

Endogenous  variables: Food  consump;on Meat  consump;on Land  use  (Cropland,  pasture,  forest)

Three  approaches  for  assuming  parameters  1.  Based  on  observed  data  (Stylized  

fact)  2.  Based  on  future  es;mates  of  

exis2ng  studies  3.  Assumed  in  line  with  SSP  

storylines  if  neither  were  available.  

Method

Parameters  related  to  food  and  hunger

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0 20000 40000 60000 80000

Coefficient  of  v

ariatio

n  (CV)  of  the

 dom

estic

 distrib

ution  of  dietary  ene

rgy  consum

ption

GDP  per  capita  [US$,  2005]

observed  data

Optimistic

Median

Pessimistic

0.134

0.109 2

0.054

0.6894

0.6531 (Observeddata; 0.61)

0.4609

y x

y x R

y x

−

−

−

= ⋅

= ⋅ =

= ⋅

Comparison  with  observed  data:  the  improved  equality  of  food  distribu;on  with  income  growth

8

Intermediate

Pessimis;c

Op;mis;c    

Equality

Inequality  (2005)

Method

Hasegawa  et  al.  in  review

0

100

200

300

400

500

600

700

800

0 20000 40000 60000 80000

Meat-­‐ba

sed  calorie

 intake  [k

cal/da

y/cap]

income  [US$/person]

observed  dataOptimisticMediumPessimistic

2

60.0 ln 399.8

87.8 ln 545.3(Observeddata; 0.57)130 ln 833.0

y x

y x Ry x

= ⋅ −

= ⋅ − =

= ⋅ −

Comparison  with  observed  data:    increased  meat  consump;on  with  income  growth

9

(1980  -­‐  2009)

Intermediate

Pessimis;c

Op;mis;c    

Method

Hasegawa  et  al.  in  review

05001000150020002500300035004000

2005

2010

2020

2030

2040

2050

2060

2070

2080

2090

2100

Carolie  intake  [kcal  /pe

rson

/day]

The  21st-­‐century  Scenarios  of  risk  of  hunger  using  SSPs  

10

Food  consump;on  per  capita

0

2000

4000

6000

8000

10000

12000

SSP1 SSP2 SSP3 SSP4 SSP5

2005 2100

Land

 use  [M

ha] Food  crops

Pasture

Grassland

Managed  forest

Primary  forest

0100200300400500600700800900

2005

2010

2020

2030

2040

2050

2060

2070

2080

2090

2100

Popu

latio

n  at  risk  of  h

unger  

[million]

SSP1SSP2SSP3SSP4SSP5

Land  use  change

Popula;on  at  risk  of  hunger

Result

0100200300400500600700800900

2005

2010

2020

2030

2040

2050

2060

2070

2080

2090

2100

Popu

latio

n  at  risk  of  h

unger  

[million]

SSP1SSP2SSP3SSP4SSP5

SSP3

SSP4

SSP2 SSP5

SSP1

Hasegawa  et  al.  in  review

Decomposi;on  analysis  What  strongly  affects  future  risk  of  hunger?

•  Change  in  hunger  risk  was  decomposed  into  three  factors;

11

Change  in  pop.  at  risk  of  hunger  

=　　　　　 　　　＋　　　　　　　　 ＋　　　　 ＋Residual            

Popula;on  growth  

Inequality  of  food  distribu;on    

Food  consump;on  

＊See  discussion  paper  for  more  detail  @hjp://www.nies.go.jp/social/dp/dpindex.html.

Method

12

Factors  affec;ng  future  hunger  risk  (global,  at  2005  level)

•  Popula;on,  inequality  of  food  distribu;on  causes  large  differences  in  hunger  risk  among  SSPs  

-­‐1500

-­‐1000

-­‐500

0

500

1000

2005 2020 2035 2050 2065 2080 2095

Change  in  pop

ulation  at  risk  of  

hunger  [m

illion]

a)  SSP1

-­‐1500

-­‐1000

-­‐500

0

500

1000

2005 2020 2035 2050 2065 2080 2095

Change  in  pop

ulation  at  risk  of  

hunger  [m

illion]

b)  SSP2

-­‐1500

-­‐1000

-­‐500

0

500

1000

2005 2020 2035 2050 2065 2080 2095

Change  in  pop

ulation  at  risk  of  

hunger  [m

illion]

c)  SSP3

-­‐1500

-­‐1000

-­‐500

0

500

1000

2005 2020 2035 2050 2065 2080 2095

Change  in  pop

ulation  at  risk  of  

hunger  [m

illion]

d)  SSP4

-­‐1500

-­‐1000

-­‐500

0

500

1000

2005 2020 2035 2050 2065 2080 2095

Change  in  pop

ulation  at  risk  of  

hunger  [m

illion]

e)  SSP5

-­‐1500

-­‐1000

-­‐500

0

500

1000

SSP1 SSP2 SSP3 SSP4 SSP5

Change  in  pop

ulation  at  risk  of  

hunger  [m

illion]

residual

per-­‐capita  calorie(domestic  production)inequality  of  food  distribution

population

net  change

f)  2100

Result

Hasegawa  et  al.  in  review

•  21st-­‐century  risk  of  hunger  differs  among  SSPs  •  Regional  distribu;on  depends  greatly  on  popula;on  growth,  equality  in  food  

distribu;on  and  increase  in  food  consump;on  •  Regions  with  greater  popula;on                  growth  face  higher  risk  of  hunger.  

Regional  popula2on  at  risk  of  hunger  and  its  factors  (SSP3,  2100）  

Hasegawa et al. in review

The  most  pessimis2c  scenario  (SSP3)  

Rest  of  Africa,  39%

India,  23%

Rest  of  Asia,  16%Middle  

East

Southeast  Asia

Rest  of  South  America

China Brazil North  Africa Former  

Soviet  Union

0100200300400500600700800900

2005

2010

2020

2030

2040

2050

2060

2070

2080

2090

2100

Popu

latio

n  at  risk  of  h

unger  

[million]

SSP1SSP2SSP3SSP4SSP5

SSP3

SSP4

SSP2

SSP5 SSP1

-­‐500  

-­‐250  

0  

250  

500  Ch

ange  in  pop

ula2

on  at  risk  of  

hunger  [m

illion]

residual  

trade  

per-­‐capita  calorie  

inequality  of  food  distribu;on  

popula;on  

(Rela;ve  to  2005)

0

200

400

600

800

1000

1200

1990 2020 2050 2080

Popu

latio

n  at  risk  of  h

unger  [million]

SSP1SSP2SSP3SSP4SSP5Alexandratos  and  Bruinsma,  2012  medArnell  et  al.,  2002  highArnell  et  al.,  2002  lowArnell  et  al.,  2002  medParry  et  al.,  2004  A1Parry  et  al.,  2004  A2Parry  et  al.,  2004  B1Parry  et  al.,  2004  B2Parry  et  al.,  2005  highParry  et  al.,  2005  lowParry  et  al.,  2005  medSchmidhuber  and  Tubiello,  2007  (AEZ+BLS)  A1Schmidhuber  and  Tubiello,  2007  (AEZ+BLS)  A2Schmidhuber  and  Tubiello,  2007  (AEZ+BLS)  B1Schmidhuber  and  Tubiello,  2007  (AEZ+BLS)  B2Schmidhuber  and  Tubiello,  2007  (DESSAT+BLS)  A1Schmidhuber  and  Tubiello,  2007  (DESSAT+BLS)  A2Schmidhuber  and  Tubiello,  2007  (DESSAT+BLS)  B1Schmidhuber  and  Tubiello,  2007  (DESSAT+BLS)  B2

•  Lines:  this  study  •  Dots:  exis;ng  studies

SSP3

SSP4

SSP2 SSP5

SSP1

Comparison  of  the  popula;on  at  risk  of  hunger  with  exis;ng  studies

14

•  Pop.  at  risk  of  hunger  in  this  study  was  lower  than  in  exis;ng  studies.  

•  Improvements  in  food  distribu;on  equality  was  considered  in  this  study  whereas  it  was  not  for  exis;ng  studies.  

•  à  Inequality  of  food  distribu2on  influences  long-­‐term  assessments  of  hunger  risk.

Result  &  discussion

Hasegawa  et  al.  in  review

Conclusion  (1/2) We  developed  scenarios  for  hunger  risk  in  the  21st  century  using  SSPs.    Factors  affec;ng  future  hunger  risk  were  inves;gated.    •  Risk  of  hunger  without  climate  change  in  the  21st  century  differed  among  SSPs  

•  Factors  affec;ng  the  future  risk  of  hunger  were  popula;on,  inequality  of  food  distribu;on,  and  per-­‐capita  food  consump;on.  –  Improving  food  access  and  consump;on  in  low-­‐income  households  will  be  effec;ve  to  reduce  future  hunger  risk.    

•  Inequality  of  food  distribu;on  greatly  influences  long-­‐term  assessments  of  hunger  risk.  

15

Conclusion

Conclusion  (2/2)

•  This  study  represents  a  star;ng  point  for  research  using  SSPs.  

•  Climate  impact  studies  an  agriculture  and  food  could  expand  by  including  addi;onal  factors.  

•  More  specific  and  tailor-­‐made  scenarios  could  be  developed  to  help  policy  decision  making  e.g.  by  based  on  stakeholders  opinions.  

16