EFFECT OF SHADE ON TEMPERATURE MITIGATION AND … · effect of shade on temperature mitigation and...

EFFECT OF SHADE ON TEMPERATURE MITIGATION AND CANOPY ASSIMILATION OF COFFEE AGROFORESTRY SYSTEMS:A MODELLING APPROACH

VEZY R *, PICART D, CHRISTINA M, SOMA M, GEORGIOU S, CHARBONNIER F, LOUSTAU D, IMBACH PB, DE MELO E., HIDALGO HG, ALFARO EJ, LE MAIRE G, ROUPSARD O

3rd European Agroforestry Conference 2016

IMPACT OF FUTURE CLIMATE ON CROP YIELD

3rd European Agroforestry Conference 2016 2

Rosenzweig, Elliott et al. (2014)

IMPACT OF FUTURE CLIMATE ON CROP YIELD

3rd European Agroforestry Conference 2016 2

Rosenzweig, Elliott et al. (2014)

First reductions will begin around 2020in African semi-arids areas, thenprogressively to south and centralamerica, Mexico, and Asia (IPCC AR5WG2).

IMPACT OF FUTURE CLIMATE ON CROP YIELD

3rd European Agroforestry Conference 2016 2

Rosenzweig, Elliott et al. (2014)

First reductions will begin around 2020in African semi-arids areas, thenprogressively to south and centralamerica, Mexico, and Asia (IPCC AR5WG2).

Adaptative options (short to long term):• Sow/Harvest dates• Agroforestry• Breeding• Crop shift

3rd European Agroforestry Conference 2016 3

Agroforestry may regulate microclimate and crop temperature.• What is the expected reduction on temperature?• Effect of shade tree species, density and management?• How to cope with local climate, elevation, bearing…• Does LUE compensate somehow for light reduction

under shade ?

A TRADE-OFF BETWEEN SHADE AND YIELD ?

Regulation of crop canopytemperature ?

LUE increaseunder shade ?

Crop yield?

3rd European Agroforestry Conference 2016 4

METHODS

Field experiments extremely useful but hard to maintain or extend:• Trees take decades to mature• Huge amount of management/climate options to test

Numeric models

Opportunity for model coupling

Agroforestry is spatialy heterogeneous

Need 3D plant-to-plot models + Accurate light interception and energy balance + process-based model + (half-) hour time step.Model selected here: MAESPA (Duursma et al. 2012)

Crop model, horizontallyuniform

Allocation, growth and yield over full rotation + Plot scale + Fast+ Process-based

MAESPA VALIDATION ON A 15 YEARS-OLD AGROFORESTRY TRIAL IN CATIE RESEARCH CENTER (Haggar et al., 2011)

3rd European Agroforestry Conference 2016 5

All credits to M.Soma

3rd European Agroforestry Conference 2016 6

MAESPA LIGHT INTERCEPTION VALIDATION (1)

Simulated Total Transmittance (Day of Year 76, by hour)

7

MAESPA light interception validation through diffuse transmittance from hemispheric photograp

MAESPA LIGHT INTERCEPTION VALIDATION (2)

3rd European Agroforestry Conference 2016

Measured diffuse transmittance (0-1)

Sim

ula

ted

DT

(0

-1)

7

MAESPA light interception validation through diffuse transmittance from hemispheric photograp

MAESPA LIGHT INTERCEPTION VALIDATION (2)

3rd European Agroforestry Conference 2016

Sim

ula

ted

CC

T (

°C)

Measured coffee canopy temperature (°C)

FUTURE CLIMATE, SIMULATED TO THE POINT

3rd European Agroforestry Conference 2016 8

Statistical downscaling from 14 GCMs to 5 km definition following Hidalgo et al. (2016, in prep)

Time (Year)

To

tal a

nn

ual

rain

fall

(mm

)M

ean

ann

ual

air

tem

p. (

°C)

Time (Year)

FUTURE CLIMATE, SIMULATED TO THE POINT

3rd European Agroforestry Conference 2016 8

Statistical downscaling from 14 GCMs to 5 km definition following Hidalgo et al. (2016, in prep)

Time (Year)

To

tal a

nn

ual

rain

fall

(mm

)M

ean

ann

ual

air

tem

p. (

°C)

Time (Year)From1979

To 2049

FUTURE CLIMATE, SIMULATED TO THE POINT

3rd European Agroforestry Conference 2016 8

Statistical downscaling from 14 GCMs to 5 km definition following Hidalgo et al. (2016, in prep)

Time (Year)

To

tal a

nn

ual

rain

fall

(mm

)M

ean

ann

ual

air

tem

p. (

°C)

Time (Year)

Aquiares, lowland,2.48°C increase for RCP8.5

FUTURE CLIMATE, SIMULATED TO THE POINT

3rd European Agroforestry Conference 2016 8

Statistical downscaling from 14 GCMs to 5 km definition following Hidalgo et al. (2016, in prep)

Time (Year)

To

tal a

nn

ual

rain

fall

(mm

)M

ean

ann

ual

air

tem

p. (

°C)

Time (Year)

Aquiares, lowland,2.48°C increase for RCP8.5

Tarrazu, mountain, 2.27°C increase for RCP8.5

FUTURE CLIMATE, SIMULATED TO THE POINT

3rd European Agroforestry Conference 2016 8

Statistical downscaling from 14 GCMs to 5 km definition following Hidalgo et al. (2016, in prep)

Time (Year)

To

tal a

nn

ual

rain

fall

(mm

)M

ean

ann

ual

air

tem

p. (

°C)

Time (Year)

Aquiares, lowland,2.48°C increase for RCP8.5

Tarrazu, mountain, 2.27°C increase for RCP8.5

Pratically no change expected for precipitations

SIMULATION OF COFFEE MANAGEMENT SCENARIOS

9

Climate Location Shade (species and density) Plot Age

RCP 4.5

X

Tarrazu

(~1500m high)

X

Full Sun: 0

X

1

2

Erythrina poeppigiana .

RCP 8.5Aquiares

(~1000m high)

Low: 200/250 High: 350/400 .

Cordia alliodora .

Low: 50/75 High: 100/125 35

3rd European Agroforestry Conference 2016

SIMULATION OF COFFEE MANAGEMENT SCENARIOS

9

Climate Location Shade (species and density) Plot Age

RCP 4.5

X

Tarrazu

(~1500m high)

X

Full Sun: 0

X

1

2

Erythrina poeppigiana .

RCP 8.5Aquiares

(~1000m high)

Low: 200/250 High: 350/400 .

Cordia alliodora .

Low: 50/75 High: 100/125 35

3rd European Agroforestry Conference 2016

Pruned twice a year to optimize coffee light intake (low LAI)

SIMULATION OF COFFEE MANAGEMENT SCENARIOS

9

Climate Location Shade (species and density) Plot Age

RCP 4.5

X

Tarrazu

(~1500m high)

X

Full Sun: 0

X

1

2

Erythrina poeppigiana .

RCP 8.5Aquiares

(~1000m high)

Low: 200/250 High: 350/400 .

Cordia alliodora .

Low: 50/75 High: 100/125 35

3rd European Agroforestry Conference 2016

Grow freely, low densities, make high coverage (High LAI)

COFFEE LIGHT USE EFFICIENCY (1979-2050 RCP 8.5, AQUIARES)

3rd European Agroforestry Conference 2016 10

COFFEE LIGHT USE EFFICIENCY (1979-2050 RCP 8.5, AQUIARES)

3rd European Agroforestry Conference 2016 10

High shade tree LAI, Coffee LUE increase

COFFEE LIGHT USE EFFICIENCY (1979-2050 RCP 8.5, AQUIARES)

3rd European Agroforestry Conference 2016 10

High shade tree LAI, Coffee LUE increaseLow tree LAI, no or little increase in LUE

COFFEE DAILY MAXIMUM CANOPY TEMPERATURE (RCP 8.5, AQUIARES)

3rd European Agroforestry Conference 2016 11

-1.96°C***, -2.58°C***-2.71°C*** -3.3°C***

CROP MODEL : Aquiares (Lowland), RCP 4.5, Shade = Cordia alliodora (50 tree ha-1, thinned)

3rd European Agroforestry Conference 2016 12

ShadetreeTransmitt-ance

Coffee LAI

Coffee canopyphotosynt.(GPP)

3rd European Agroforestry Conference 2016 13

Coffee yieldwithshadetrees

CROP MODEL : Aquiares (Lowland), RCP 4.5, Shade = Cordia alliodora (50 tree ha-1, thinned)

Air T (°C)

Infloresce-ncesper node

CONCLUSION

• Coupling a 3D with a crop allocation model succesfully simulates leaf temperature, photosynthesis, light use efficiency and yield over full rotations

3rd European Agroforestry Conference 2016 14

CONCLUSION

• Coupling a 3D with a crop allocation model succesfully simulates leaf temperature, photosynthesis, light use efficiency and yield over full rotations

3rd European Agroforestry Conference 2016 14

• Coffee yield could be effectively sustained under future climate through shade management

CONCLUSION

• Coupling a 3D with a crop allocation model succesfully simulates leaf temperature, photosynthesis, light use efficiency and yield over full rotations

3rd European Agroforestry Conference 2016 14

• Coffee yield could be effectively sustained under future climate through shade management

• Models should allow to optimize shade and yield under various geographic location, management and climate scenarios

CONCLUSION

• Coupling a 3D with a crop allocation model succesfully simulates leaf temperature, photosynthesis, light use efficiency and yield over full rotations

3rd European Agroforestry Conference 2016 14

• Coffee yield could be effectively sustained under future climate through shade management

• Models should allow to optimize shade and yield under various geographic location, management and climate scenarios

• Simulations will be extended to 2100 (no strong RCPs differences until 2049)

CONCLUSION

• Coupling a 3D with a crop allocation model succesfully simulates leaf temperature, photosynthesis, light use efficiency and yield over full rotations

3rd European Agroforestry Conference 2016 14

• Coffee yield could be effectively sustained under future climate through shade management

• Models should allow to optimize shade and yield under various geographic location, management and climate scenarios

• Simulations will be extended to 2100 (no strong RCPs differences until 2049)

• Further analyses will be conducted over all the scenarios

3rd European Agroforestry Conference 2016 15

• ANR MACACC Project

• INRA

• CIRAD

• UMR ECO&SOLS / UMR ISPA

• CATIE

• UCR (University of Costa Rica)

ACKNOWLEDGMENTS

ANNEXES

3rd European Agroforestry Conference 2016 16

NUMERIC MODELS SCHEME

17

Meteorologicalvariables

MAESPA Outputs

Meteorologicalvariables

Allocation model

Parameters:Structure,

physiology, soil…

Metamodel: LUE & 𝐾𝐷𝑖𝑓𝑓𝑢𝑠 &𝐾𝐷𝑖𝑟𝑒𝑐𝑡

Parameters:Structure,

physiology, soil…

Outputs (=management

simulations)

Input parameters (as is or modified)

3rd European Agroforestry Conference 2016

NUMERIC MODELS SCHEME

17

Meteorologicalvariables

MAESPA Outputs

Meteorologicalvariables

Allocation model

Parameters:Structure,

physiology, soil…

Metamodel: LUE & 𝐾𝐷𝑖𝑓𝑓𝑢𝑠 &𝐾𝐷𝑖𝑟𝑒𝑐𝑡

Parameters:Structure,

physiology, soil…

Outputs (=management

simulations)

Input parameters (as is or modified)

3rd European Agroforestry Conference 2016

DOWNSCALING TECHNIQUE

• Statistical downscaling:

– Depends on current empirical relationships

– Low computational demand

– Long and high quality data series

– Hard to apply in complex environments