The document summarizes the results of a workshop on climate change impacts on cocoa in Ghana. Random forest machine learning was used to analyze climate and soil data and classify suitable cocoa growing zones. Four key zones were identified: 1) low rainfall, long dry season, poor soils in northwest; 2) low temperatures, long dry season, average soils; 3) high temperatures, reliable rainfall, average soils; and 4) high temperatures, low seasonality, rich soils. Analysis found temperatures in some current cocoa areas may exceed tolerable limits by 2050, especially in northwest Ghana. Soil characteristics will influence resilience to climate impacts. The moist evergreen zone may become most suitable while the northwest may become marginal.
5. International Center for
Tropical Agriculture, CIAT
• 50 years of applied research
for improved livelihoods and
environmental sustainability
in the global tropics.
• 900 staff active in Africa,
Latin America and South
East Asia.
• Annual budget of US 130m.
• Lead center for the global
Climate Change, Agriculture
and Food Security Program
of the CGIAR.
6. International Center for
Tropical Agriculture, CIAT
Role in this project
• Mapping risk of climate
change for cocoa in Ghana
• Economic analysis of cost
and benefits of adaptation
strategies
• How to scale CSA practices
in cocoa systems
• Overall project and
consortium management,
reporting and learning.
7. International Institute
of Tropical Agriculture
One of the world's leading research institutes working
with partners in Africa and beyond to reduce producer
and consumer risks, enhance crop quality and
productivity, improve livelihoods and generate wealth
from agriculture.
8. International Institute
of Tropical Agriculture
Project role
• Coordination in Ghana together with RA
• Situational analysis
• Stakeholder engagement
• Social learning
• Identify strategic learning sites along climate gradients
• Develop relevant adaptation practices for cocoa
• Climate Smart Agriculture planning that fosters gradual
change/transition in the identified high impact zones
• Match CSA to value chain actors’ needs according to the
agreed identified adaptation zones
11. Project objectives
The project expects to contribute to:
Clear knowledge of what types of CSA practices to promote
where, for whom and with what return on investment
Knowledge of under what conditions extension and PO
investments function as incentives for CSA uptake at scale
Identification of additional public, private or public-private
incentives needed to promote widespread CSA adoption in the
cocoa sector
Functional multi-stakeholder platforms that combines climate
science with industry knowledge to reduce risk faced by cocoa
in Ghana going forward.
• We seek to add value to what all of you are already
doing around climate change and look forward to
hearing what you think, how we might best collaborate
and what additional issues should be considered.
13. Previous studies
• Suitability
losses in the
West
• Some gains
towards Lake
Volta.
Läderach et al. (2013)
Predicting the future climatic
suitability for cocoa farming of
the world’s leading producer
countries, Ghana and Côte
d’Ivoire” Climatic Change.
14. Previous studies
• Ghana:
Losses in the
North, Gains in
central areas.
• West-A.:
Maximum dry
season
temperatures
seen to be
problematic.
• West-A.: Areas
at the margins
to Savanna
are most
vulnerable.
Schroth et al.,“Vulnerability to climate change of cocoa in West Africa: patterns,
opportunities and limits to adaptation” Agriculture, Ecosystems and Environment (Submitted).
15. Previous studies
• Some disagreement about distribution of impacts
• Unspecific to Ghana
Which are the climate events Ghana has to prepare for?
• „Suitability“ = Probabilities from binary classification method
How much „suitability loss“ is critical?
What can be done to adapt?
16. Application of a machine learning tool to cocoa in Ghana
Random Forest classification
17. When to apply machine learning?
Where can we grow cocoa?
„For optimal conditions maximum temperatures should
not exceed 32 °C (Lass and Wood 1985)
Not in Ghana!
• Our understanding of climatic
requirements is often limited
• Our climate data is often bad
• Crop simulation models are
very complex
• Crop simulation models often
give unsatisfactory results even
for rice and maize
18. Random Forests for classification
• A forest is an ensemble of
trees. The trees are all
slightly different from one
another.
• The output is the mean
classification
• Very robust against
overfitting
Is the soil good?
Is the dry
season long?
Is the heat
strong?
One decision tree
Source: Criminisi et al 2013
19. Random Forest classification
• Training classes:
5 AEZ clusters as
suitable classes
Random sample
from the area of
Ghana
Balanced
subsample
• Climate variables:
20 bioclimatic
variables
• 25 repeats
Different
subsamples
Random repeats
• 1000 trees grown,
4-5 variables picked
Type
Bioclimatic
variable
Description
Current
Mean
2030s
Mean
2050s
Mean
Unit
Temperature
BIO 1 Annual Mean Temperature 26,2 27,3 27,7 °C
BIO 2
Mean Diurnal Range (Mean of
monthly (max temp - min temp))
9,1 8,7 8,6 °C
BIO 3 Isothermality (BIO2/BIO7) (*100) 72 70 70 -
BIO 4
Temperature Seasonality (standard
deviation *100)
103,7 110,1 109,8 °C
BIO 5
Max Temperature of Warmest
Month
33,1 34,1 34,5 °C
BIO 6 Min Temperature of Coldest Month 20,6 21,9 22,3 °C
BIO 7
Temperature Annual Range (BIO5-
BIO6)
12,5 12,3 12,2 °C
BIO 8
Mean Temperature of Wettest
Quarter
26,5 27,2 27,6 °C
BIO 9 Mean Temperature of Driest Quarter 26,6 27,8 28,2 °C
BIO 10
Mean Temperature of Warmest
Quarter
27,4 28,7 29,1 °C
BIO 11
Mean Temperature of Coldest
Quarter
24,7 25,8 26,2 °C
Precipitation
BIO 12 Annual Precipitation 1453 1463 1476 mm
BIO 13 Precipitation of Wettest Month 234 233 235 mm
BIO 14 Precipitation of Driest Month 22 21 21 mm
BIO 15
Precipitation Seasonality (Coefficient
of Variation)
53 54 55 -
BIO 16 Precipitation of Wettest Quarter 570 567 575 mm
BIO 17 Precipitation of Driest Quarter 117 114 113 mm
BIO 18 Precipitation of Warmest Quarter 335 326 329 mm
BIO 19 Precipitation of Coldest Quarter 361 379 382 mm
BIO 20
Number of Consecutive Months <
100mm precipitation
3,63 3,65 3,63 -
25. Clustering result
1 Elevated temperatures Reliable precipitation Average soils
2 Low annual precipitation Strong dry season Below average soils
3 High temp Low seasonal variation Above average soils
4 Low temperatures Long dry season Average soils
27. Current distribution of suitability classes
for cocoa
• MSNW
Moist semi-decidious North
West
• MSSE
Moist semi-decidious South-
East
• ME
Moist evergreen
30. Current distribution of suitability classes
for cocoa
AEZ Bioclim A Bioclim B Soils
Type 1 Low annual precipitation Strong dry season Below average soils
Type 2 Low temperatures Long dry season Average soils
Type 3 Elevated temperatures Reliable precipitation Average soils
Type 4 High temp Low seasonal variation Above average soils
38. Conclusion
• Cocoa production is shaped by climate and soils
• Cocoa soil characteristics are different from other soils in the
country
• Results show four distinct production zones that align with
ecological zones in Ghana:
Moist semi-deciduous, subtypes NW, (central),SE
Moist evergreen
• The moist evergreen climate (type 4) will be the dominant climate in
the future
• The moist semi-deciduous (type 1) region in the North West will
become marginal
• Soils will determine the resilience against climatic change
39. Acknowledgements
• Sander Muilerman (IITA)
• Christian Mensah (Rainforest
Alliance)
• Dr. Anim-Kwapong (CRIG)
• Dr. Amos Quaye (CRIG)
• Patrick Adjewodah (IITA/RA)
• Workshop participants from CRIG:
E. Amamoo-Otchere
Patrick Adjewodah
A. Afrifa
Godfrend Awudzi
Robert Asugre
Jerome Dogbatse
Dr. Sampson Kolan
Fredrick Amon-Armah
Esther Gyan
Williams Atakorah
Mustapha Alasan Dalaa (IITA)