Trees and Dietary Diversity in Africa Amy Ickowitz September 23, 2013 Presentation for Science Forum 2013, Bonn, Germany
Jul 08, 2015
Trees and Dietary Diversity in AfricaAmy Ickowitz
September 23, 2013
Presentation for Science Forum 2013, Bonn, Germany
Hypothesis: Trees and Forests are important for dietary quality (dietary diversity)
� Collection of nutritious NTFPs
� Farming mosaics may promote
more diverse diets
� Agroforestry and fruit
production
� Ecosystem services of forests for
agriculture
� Availability of fuel wood
� May provide ‘back up’ foods for
lean season
• Forests may be relevant for food security/nutrition discussion not
only because of potential conflicts with another SLO, but because
might be of direct importance to nutrition as well
• Choices that are framed as food security/nutrition or conservation
are more complex (especially relevant for govt policy)
• We may need to fundamentally rethink the way we do our
research ….title of this Science Forum
Some implications, if true
Is there evidence?
� Several recent papers discussing some of these plausible links
(Colfer et al., 2005; Vinceti et al., 2008; Arnold et al., 2011), but little
data to support
� Recent paper by Johnson et al. (2013) finds that net forest loss
associated with less dietary diversity in Malawi
� Also, it is plausible that there might be a negative association
• Study using DHS data from 21
countries integrated with GIS
data on % tree cover to
estimate the relationship
between tree cover and child
nutrition indicators (Ickowitz,
Powell, Salim, Sunderland,
under review)
• CIFOR project collecting dietary
intake information from
mothers and children in study
sites in 5 African countries
Testing the hypothesis
We Integrate:
• nutrition data from Demographic Health Surveys
with
• % tree cover data from GLCF (2003 and 2010
MODIS data at 250 m resolution)
(as well as other sources for other controls)
to investigate whether there is a statistically
significant relationship between indicators of
dietary quality and tree cover
Study using DHS data
Sample: about 93,000 children between ages 13 and 59 months in over 9,500 communities
(21 countries )
• 3 Dependent Variables:
DDS; Fruit & Vegetable consn; Animal Source
Food consn
• Independent Variables:� % Tree Cover and % Tree Cover2
� Mother’s education
� Father’s education
� Wealth Index
� Rural Dummy
� Child age & age2 & age3
� Distance to Rd
� Distance to closest city of 10,000
� Aridity Index
� Elevation
� Sex of child
� Currently breastfeeding
� Month of interview
� Country dummy
Regressions
• There is a statistically significant positive relationship between % tree
cover and Dietary Diversity (as proxied by DDS)
• Fruit and Vegetable Consumption first increases and then decreases
with tree cover (peak tree cover is about 45%)
• There is no statistically significant relationship between tree cover
and Animal Source Foods
Results
• The results of the DHS study give
an indication that there are
interesting relationships, but are
far from offering an explanation
• DHS data are coarse
• The GIS data don’t tell us the
kinds of trees
• Data can’t explain WHY
children in areas with more
trees have more diverse diets
• Country level regressions give
heterogenous results
• So….
Where are we now?
• Collect data with more detailed
dietary info. and on where the food
actually comes from (forests,
fallows, agro-forests, farms, etc.)
• Compare diets in villages with
varying % of tree cover in
different regions of Africa:
Uganda, Ethiopia, Cameroon,
Zambia, Burkina Faso
• Two year project
• Already begun; preliminary
results available by end of
2014
Nutrition and trees in sub-Saharan Africa: A Regional Comparative Study
� We have found a statistically significant relationship between
tree cover and dietary diversity in a large sample of African
children
� This gives preliminary support to the hypothesis that
forests are important for nutrition in Africa
� We need more detailed and fine-grained data to help us to
understand how and why children in more forested areas
in Africa have higher dietary diversity
� We hope that our findings, in conjunction with recent findings
from Bioversity, WorldFish, USAID, and ICRAF can help to
convince the CG to look at the wider landscape (in addition to
agriculture) in thinking about nutrition and health
Conclusions
THINKING beyond the canopy
Influencing the agenda?
• Discussion Paper: starting point for engagement
• Representation at numerous international events (FAO, Leverhulme, IUNS, CGIAR Science Meeting)
• Collaboration with CRP4 –flagship proposal on “nutrition-sensitive landscapes”
• Clear contribution to SLO’s
• Engaging with the wider CGIAR!
THINKING beyond the canopy
Thank you!