correlated extremes for agriculture and food security ...€¦ · outline 2 agriculture and food...

Correlated Extremes for

Agriculture and Food Security ApplicationsAlex Ruane, Climate Impacts Group

NASA Goddard Institute for Space Studies, New York

Columbia University Center for Climate Systems Research

Correlated Extremes WorkshopMay 30, 2019

Presentation represents author, not necessarily NASA

National Aeronautics and Space Administration Goddard Institute for Space Studies

Goddard Space Flight Center Sciences and Exploration Directorate

Earth Sciences Division

clJ COLUMBIA

https://ntrs.nasa.gov/search.jsp?R=20190026592 2020-08-03T13:31:22+00:00Z

Outline2

Agriculture and Food Security are vulnerable to

correlated extremes

Compound Extremes

Concurrent Extremes

Sequential Extremes

Complex nature of food systems can exacerbate or

buffer ‘shocks’

Designing models and simulations to explore extreme

responses

Bonus opportunities for contribution (IPCC and GRAF)

Correlated Extremes in the Agricultural Sector

and Broader Food System

4

5

The Complex Food System

6http://www.nourishlife.org/teach/food-system-tools/

rr, :::, < -, 0 ::,

3 (t) :, rt' Q) ,-

IIOLOGICAL SYSTEM

N.da,,l

SOCIAL SYSTEM

Socia\

ECONOMIC SYSTEM

E 0 C: 0 c..,

Lu

7

Maize Production (1000s of kg)

Top Regions Accounting for 90% of World Maize Production

Tracking a Global Commodity

Data from Monfreda et al., 2002

0 5 10 15 20 25 30 0 5 10 15 20 25 30

Correlated Extremes for Food Systems

8

Compound Extremes:

Hot/wet: Livestock and labor productivity; Disease pressure

Hot/dry: Crop production, water resources

Hot/Ozone: Crop production

Concurrent Extremes:

Multi-breadbasket failure

Crop production loss

and interruption of market access

Sequential Extremes:

Early warm spell followed by frost

Early season wet spell followed by

late season drought

Food Security:

Availability, Access, Utility, Stability

The Complex Food System

9

Assumption of perpetually declining food prices

now called into question.

Spikes due to: drought, heat waves, and wildfires causing poor production in some regions;

energy price spikes, declining food stocks, trade policy, expansion of biofuels

Consequences: increased number of malnourished, shift in diets, reduced spending on

other essentials, gender-based outcomes, social unrest, migration

FAQ Food Price Index in nominal and real terms

2002-2004= 100

250

Food price spikes can exacerbate instability

10

Vulnerability:Urbanconsumers

280

260

240

220

X Q)

"tJ 200 C

Q) u '-c..

"tJ 180 0 0

I.I.

160

140

120

2004

Algeria (8), Saudi Arabia {15) 0 (2) M (l0) H ·t · ( ) E t (3) Mauritania {3), Suda n (2), Jordan (4) 1 ma(n35) •8 horo.cco(98)

a1 1 5, gyp . Yemen (2 000+ / raq , a rain Cote d' Ivoire (1) Somali~ (5) ' / / Syria (20,000+)

1

) Tuni sia (1) Egypt (800+) 1

// Uganda (5), Ira n {12) Sudan (3)i ( Libya (30,000+) ,..,.- I~ Georgia (2), Israel (30)

Cameroon (40) T . . (300 ) : : - Kenya (2), Malawi {18) Yemen (12) l unisia + - ( :: ·-- soma lia (10)

Mozambique {6) / :: India (1) ~::: Sudan (1) ~:::

Burundi (1)

2006

... , Mozambique (13) ~:::

Mauritania (2) , •••• ~-·· India (4) :

Somalia (5) : : I

I I I I

I I I I I I I I

I I I I I I I I I I

I I

I I

I I I I

I I I I

I I I I I I I I

I I

2008

.. , •.. , ~-11 .. , •••• ~-·· ••• ... , ~-·· .. , ... , ~-11 .. , ... , ~-·· ••• .. , ~-·· .. , ... , ~-·· .. , ... , ~-·· .. , .. , ~-·· .. , ••••

2010 2012 2014

Red dashed lines correspond to t he beginn ing dates of"food riots'' and protests in North Africa and the Midd le East between 2004 and 2011. The overall death toll is indicated in parent heses next to each country.

Source: Lagi, Bertand, Bar-Ya m 2011.

11

The Agricultural Model

Intercomparison and

Improvement Project (AgMIP)

12

Modified from Rosenzweig et al., 2013 AgForMet

AgMIP Enables Evaluation of Agricultural Risk

and Testing of Risk Reduction Strategies

The Agricultural Model Intercomparison and Improvement Project (AgMIP)

was created in 2010 to provide a community for systematic improvement

and application of multi-disciplinary, multi-model, multi-scale frameworks

for agricultural development and food security.

Historical

Climate

and Policy

Future

Climate

and Policy

Agricultural ModelsPhysical, biophysical,

and socio-economics

Ag M I P The Agricultural Model lntercomparison and Improvement Project

Track 1

Track 2

Mod el calibration and improvement

Adaptation, mitigation, and extensions

Evaluation and

iratertcomparison

Future agricultural production, trade, and food security

Ozone and Air Pollution

Climate Scenarios

AgMERRA

Impacts Explorer

Seasonal Forecasting

Soy

Current AgMIP ActivitiesBuilding an integrated framework across scales, disciplines, and models

13

AgMIP is an international community

of 1000+ climate scientists,

agronomists, economists,

and IT experts working to improve

assessments of future food security

Visit www.agmip.org

for more information

Process-based Cropping Systems Models Capable

of Compounding Impacts Across Time and Hazards

CO2Light Temperature

Rainfall /

Irrigation

Crop Management

Carter 2013

SoilTime (daily steps)

Ozone

Cultivar Agronomy

Breeding

Slide courtesy of Senthold Asseng, UFlorida

15

Applications Across Time HorizonsCan we find/generate/respond to correlated extremes?

Ruane et al., in prep

Ag M I P The Agricultural Model lntercomparison and Improvement Project

Cyclones Pests & Diseases Air Pollution

Detection and Attribution

Cou nterfactua I Management

Historical

Retrospective Analysis

& Extreme Storms

"::.<:'"'- , , , .. ,,, ,

::,O}.,., ... ~ , , ~

p , ~ ~~ o:: :.o.. - • • • • Temperature Extremes Rainfall Extremes

Real-time and Seasonal Outlook

Monitoring Forecasting

• Climate

' ' . Change

- Market -~ ~::,, Influences ~

>,_ Policy Change

Socioeconomic Change

'...Jit. Environmental 111111 Sustainability

Long-term Outlook

Projections

Lead Time

Designing Experiments to Explore

Unobserved Agro-Climate Conditions

16

The AgMIP

Global Gridded Crop Model

Intercomparison

(GGCMI Phase 2) examines

CO2-temperature-water-

nitrogen-adaptation

sensitivity tests across

multiple crops, models, and

farm systems

Impacts on yield, water use,

and nutritional quality

[CO2] = 360, 510, 660, 810 ppm

ΔT = -1 +0 +1 +2 +3 +4 +6 ⁰C

ΔW = -50% -30% -10% +0% +10% +30%

N = 10, 60, 200 kg N/ha

A = Fixed cultivars,

Cultivars selected to restore growing season length

[CO2]

ΔT

Mean Yield response (% change) from 106

rain-fed maize sites around the world

Mean Temperature Change (⁰C)

[CO

2]

(pp

m)

Colors = emulated mean yield changes

Gray contours = uncertainty (lighter is more uncertain)

= current conditions

GG

CM

I P

ha

se

2

17

1988 Drought

pDSSAT

crop model

Ag MIP

. , .. ,.

' ~ .. .

The Agricultural Model lntercomparison and Improvement Project

. ...

.. ... .....

-50 -40 -30

~ -.

. ... . .... .........•

-20 -10 0 10 20

percent change from C36O, TO, WO, N2OO 30 40

··• . .... .

r, . ···• ·, ....

50

18

Plausible Events and

Climate Change-driven shifts in Risk Profile

19881988 – 1ºC 1988 + 1ºC

1988 + 10% rain

1988 – 10% rain

How would 1988 drought

have been different if it were:

Wetter

Warmer

pDSSAT

crop model

19

Many individual years evaluatedMaize Yield by Climate Anomaly in Corn Belt, USA

Corn

Belt M

aiz

e P

roduction A

nom

aly

(%

)

Models uniquely

capable of exploring

unobserved

climate extremes,

field environments,

genetics, and

farm management

to characterize

particular

vulnerabilities

Filled contours show smooth fit to seasonal production anomaly as function of mean T, P

Dots represent individual years influenced by correlated extremes and phenological timing

The Intergovernmental Panel on

Climate Change Sixth

Assessment Report (IPCC AR6)

and

The UNDRR Global Risk

Assessment Framework

IPCC Sixth Assessment Report – Enhanced Focus on Risk

21

Working Group 1:

‘Regional Chapters’

Chapter 10: Linking global to regional climate change

Chapter 11: Weather and climate extreme events in a changing climate

Chapter 12: Climate change information for regional impact and for risk

assessment

• Risk = hazard x vulnerability x exposure

• Impact = a specific realization of a risk

• A climate value becomes a climate hazard when it connects to exposure

and/or vulnerability

• Hazards can change in their magnitude, frequency, duration, timing, and/or

spatial extent

Submit by December 31st, 2019!

UNDRR Global Risk Assessment Framework

22

The UNDRR GRAF

aims to provide a

curated space for

disaster systems

assessment, providing

an infrastructure to

foster improved risk

information for

stakeholders

Target: 2030

Check out GAR19

Summary

23

Vulnerabilities to correlated extremes are

increasingly recognized within the agricultural

sector and broader food system

Compound, concurrent, and sequential extremes

pose unique challenges, particularly when coupled

with field management and markets that span

many spatial scales and time horizons

AgMIP is systematically connecting climate to crop,

livestock, economics, food security, and nutrition

using a coherent modeling framework that captures

more complex behaviors of correlated extremes

Experiments designed to explore historical (as well

as hypothetical) extremes aid in big data analysis

and planning for resilience and broader risk

management

Opportunities to better connect extremes, hazards,

disasters and climate through IPCC and GRAF

Thanks! alexander.c.ruane@nasa.gov