Simulation of tea yield with AquaCrop...Apr 30, 2013  · unit ground surface soil surface covered...

Simulation of tea yield

with AquaCrop Dirk RAES, KU Leuven University, Leuven Belgium Patricia MEJIAS, FAO, Land and Water Division, Rome, Italy Samson KAMUNYA, John BORE, Beatrice CHESEREK, Paul KIPRONO, David KAMAU, Tea Research Foundation of Kenya Aziz Elbehri, FAO, Land and Water Division, Rome, Italy

1

Climate Change and the Tea Sector in Kenya: Impact Assessment and Policy Action

National Multi-stakeholder Workshop 29-30 April 2013, Naivasha

2

1. Calculation scheme

Crop development

Crop transpiration

Biomass production

Yield formation

Structure of the presentation

2. Running simulations

Experimental fields

Farmer’s fields

Effect of climate change

3

Instead of Leaf Area Index (LAI)

AquaCrop uses green canopy cover (CC)

CC = soil surface covered by the green canopy

unit ground surface area

ranges from 0 (bare soil) to 1 (full canopy cover)

0 % 100 %

unit ground surface

soil surface covered by green canopy

CC = 25 %

CC = 95 %

Lung pruning

4

Canopy development (non-limiting conditions)

1.5 years

4 years

CCini = 25%

CCx = 95%

pruning 5

Relative tea yield (non-limiting conditions)

pruning

1st

year

2nd

year

3rd

year

4th

year

6

7

irrigation (I)rainfall (P)

capillary

rise deep

percolation

sto

red

so

il w

ate

r (m

m)

field capacity

threshold

wilting point

evapo-

transpiration

(ET)

(CR)

(DP)0.0

Water stress (upper) thresholds

FC

PWP

leaf expansion

canopy senescence

0.10 – 0.40 TAW

0.60 – 1.00 TAW

TAW

Effect of water stress on canopy development

1 Jan 1996

31 Dec 1997

slow expansion

senescence

8

9

Transpiration = Kssto x KcTr x ETo

no water stress

reference evapotranspiration

crop coefficient

weather conditions

characteristics of the transpiring crop

evaporative power of the atmosphere

proportional factor (KcTr,x) = 0.85 (integrating the effects of characteristics that distinguish the crop from the reference grass)

proportional to green canopy cover (CC)

water stress coefficient for stomatal closure

10

irrigation (I)rainfall (P)

capillary

rise deep

percolation

sto

red

so

il w

ate

r (m

m)

field capacity

threshold

wilting point

evapo-

transpiration

(ET)

(CR)

(DP)0.0

Water stress (upper) thresholds

FC

PWP

leaf expansion

stomatal closure

canopy senescence

0.10 – 0.40 TAW

0.60 – 1.00 TAW

0.25 – 1.00 TAW

H O2

CO2

11

1. Calculation scheme Crop development

Crop transpiration

Biomass production

CC

WP*

WP* normalized biomass water productivity

12

13

1. Calculation scheme Crop development

Crop transpiration

Biomass production

CC

WP* 14 g/m2

bi = Ksb WP* (Tri/EToi)

Tbase = 8°C cold stress

Tmax

Tmin

Effect of cold stress on biomass production

°C

month

14

15

1. Calculation scheme Crop development

Crop transpiration

Biomass production

Yield formation

CC

WP*

HI 14 %

14 g/m2

30 % = 1.2 years

14 %

16

17

1. Calculation scheme Crop development

Crop transpiration

Biomass production

Yield formation

CC

WP*

HI 14 %

14 g/m2

yi = Ksexp,w HI bi water stress

Rain

water stress 18

2005 2006

Effect of water stress on tea yield

19

20

1. Calculation scheme

Crop development

Crop transpiration

Biomass production

Yield formation

Structure of the presentation

2. Running simulations

Experimental fields

Farmer’s fields

Effect of climate change

Experimental fields

WP* = 14 g/m2

HI = 14 % CCx = 95 %

1 2 3 4 1 2 3 4 ….

21

pruning pruning

observed tea yield

simulated tea yield 22

Experimental fields

WP* = 14 g/m2

HI = 14 % CCx = 95 %

Farmer’s fields

WP* = 14 g/m2

HIo = 10 % CCx = 70 %

1 2 3 4

1 2 3 4

1 2 3 4

1 2 3 4 …. ….

…. ….

….

23

1 2 3 4 1 2 3 4 ….

pruning pruning

average observed tea yield

average simulated tea yield 24

average simulated tea yield

1st year

3rd year average observed tea yield

25

26

Statistical

indicator

Value Observation

R2 0.76 values greater than 0.50 are considered acceptable

RMSE 261 kg/ha it summarizes the mean difference in simulated and

observed

NRMSE 10.0 % a simulation can be considered excellent if NRMSE

is smaller than 10%

EF 0.66 an EF of 1 indicates a perfect match between the

model and the observations, an EF of 0 means that

the model predictions are as accurate as the average

of the observed data

d 0.85 0 indicating no agreement and 1 indicating a perfect

agreement between the predicted and observed

data. This statistical indicator overcomes the

insensitivity of R² and EF to systematic over- or

underestimations by the model

27

28

Effect of climate change Crop development

Crop transpiration

Biomass production

Yield formation

CC

WP*

HI

less water stress less temperature stress

+ + + +

CO2 fertilization

+++

heat stress ?

-