Content uploaded by Joel A. Carr
All content in this area was uploaded by Joel A. Carr on Oct 20, 2015
Content may be subject to copyright.
Feeding humanity through global food trade
Paolo D’Odorico1, Joel A. Carr1, Francesco Laio2, Luca Ridolﬁ2, and Stefano Vandoni2
1Department of Environmental Sciences, University of Virginia, Charlottesville, Virginia, USA, 2Dipartimento di
Ingegneria Ambientale, DIATI, Politecnico di Torino, Turin, Italy
Abstract The recent intensiﬁcation of international trade has led to a globalization of food commodi-
ties and to an increased disconnection between human populations and the land and water resources
that support them through crop and livestock production. Several countries are not self-suﬃcient and
depend on imports from other regions. Despite the recognized importance of the role of trade in global
and regional food security, the societal reliance on domestic production and international trade remains
poorly quantiﬁed. Here we investigate the global patterns of food trade and evaluate the dependency of
food security on imports. We investigate the relationship existing between the trade of food calories and
the virtual transfer of water used for their production. We show how the amount of food calories traded in
the international market has more than doubled between 1986 and 2009, while the number of links in the
trade network has increased by more than 50%. Likewise, global food production has increased by more
than 50% in the same period, providing an amount of food that is overall suﬃcient to support the global
population at a rate of 2700– 3000kcal per person per day. About 23% of the food produced for human
consumption is traded internationally. The water use eﬃciency of food trade (i.e., food calories produced
per unit volume of water used) has declined in the last few decades. The water use eﬃciency of food pro-
duction overall increases with the countries’ aﬄuence; this trend is likely due to the use of more advanced
Global food security depends on a number of rapidly changing factors that control the availability of
and access to food commodities. In the last three decades research on food security has emphasized the
importance of factors controlling the access to food [e.g., Sen, 1981]. More recently, however, scientists
have started to focus again on food suﬃciency and availability [e.g., Foley et al., 2011; Tilman et al., 2011;
Cassidy et al., 2013]. There is a growing concern that the limited resources of the planet will soon limit our
ability to keep up with the increasing food demand by human societies [e.g., Godfray et al., 2010]. While
the demand for food is steadily increasing and is competing with the rapidly growing need for agricultural
products by other sectors such as biofuel production [e.g., Godfray et al., 2010], the supply of food crops
is reaching a plateau [Ray et al., 2012]. On the demand side, demographic growth and changes in diet are
placing unprecedented pressure on the food production system [e.g., Tilman et al., 2011; Cassidy et al.,
2013; Hermele, 2014]. The global population has increased by 1 billion every 12– 14 years since the 1960s
[U.N. Population Division, 2012] and has been paralleled by an increase in food availability resulting from
the use of industrial fertilizers, irrigation technology, and new cultivars [Boserup, 1981; Erisman et al., 2008].
Interestingly, food consumption has not grown linearly with the population size as the per capita demand
for food has increased. The portion of the diet consisting of fats and proteins tends to increase with the
economic development of emerging countries [Delgado, 2003], a phenomenon known as “Bennett’s
law” [e.g., U.K. Oﬃce for Science, 2011]. In particular, the consumption of meat and other animal products
has been reported to increase with the per capita gross domestic product (GDP) and the gross national
income (GNI) [Tilman et al., 2011; U.K. Oﬃce for Science, 2011]. Because several kilocalories of forage or
feed are needed to produce 1 kcal of meat [Pimentel et al., 1973], this increase in the consumption of ani-
mal products is further enhancing the human pressure on croplands and rangelands [Cassidy et al., 2013].
On the supply side, after the extraordinary increase in crop yields in the second half of last century (i.e.,
during the “green revolution”), the rates of agricultural production are expected to reach a plateau [Ray
et al., 2012; Brown, 2013]. Likewise, it has been argued that capture marine ﬁsheries might be at the verge
of a sudden decline [Coll et al., 2008], while the ongoing increase in ﬁsh production from aquaculture, is
• We reconstruct the global network
of food trade
• We determine the dependency of
global access to food on
• We evaluate the water use eﬃciency
of food production and trade
P. D’Odorico, email@example.com
and S. Vandoni (2014), Feeding
humanity through global food trade,
Received 16 APR 2014
Accepted 5 AUG 2014
Accepted article online 12 AUG 2014
This is an open access article under
the terms of the Creative Commons
License, which permits use and distri-
bution in any medium, provided the
original work is properly cited, the use
is non-commercial and no modiﬁca-
tions or adaptations are made.
D’ODORICO ET AL. © 2014 The Authors. 1
partly sustained by feed inputs from terrestrial sources [Duarte et al., 2009; Gephart et al., 2014]. Moreover,
recent U.S. and E.U. policies have mandated the use of a prescribed fraction of energy from renewable
sources [EISA, 2007; E.U., 2009]. The consequent escalating use of crops for biofuels has started to com-
pete with food production [Hermele, 2014]. Further, the intensiﬁcation of climate extremes is predicted
to increase the variability and uncertainty of crop yields [IPCC, 2013], thereby contributing to food price
volatility [FAO-OECD, 2011]. The picture emerging from these analyses is that, in the near future, the rates
of global food supply might not meet the escalating demand.
At smaller scales, however, many regions are already in conditions of food shortage due to a local imbal-
ance between production and demand (see Supporting Information, Figure S1) [Fader et al., 2013; Suweis
et al., 2013]. The food security of these societies relies on the importation of food products from other
regions. It has been argued that the intensiﬁcation of international trade has been one of the major fac-
tors contributing to changes in food supply in the past few decades [D’Odorico and Rulli, 2013]. Global
trade has been found to account for 12% of the human appropriation of the net primary productivity and
24% of the global ecological (i.e., land) footprint [Erb et al., 2009; Weinzettel et al., 2013; Meyfroidt et al.,
2013]. The extent to which global food security depends on international trade, however, remains poorly
understood [Porkka et al., 2013]. It is unclear which region of the world is most beneﬁting from the ongo-
ing intensiﬁcation of food trade and how trade diﬀerentially aﬀects food security in developed, emerging,
or developing countries [MacDonald, 2013].
There is also some concern that food production might not grow indeﬁnitely due to limitations imposed
by land and water resources. Because of the limited potential for sustainable expansion of agricultural
land, for most regions of the world, an increase in food production requires the closure of the yield gap
in underperforming agricultural lands with strong potential for increased yields [Foley et al., 2011; Mueller
et al., 2012]. With nitrogen fertilizers becoming available in almost “unlimited” amounts through industrial
synthesis [Erisman et al., 2008], water and suitable land remain the major limiting factors for food produc-
tion. It is therefore important to evaluate the water use eﬃciency (WUE) of food commodities (kcal/L) and
investigate whether trade contributes to a more eﬀective use of freshwater resources.
Here we investigate the global patterns of food trade, identify regions of self-suﬃciency and trade depen-
dency, and evaluate how the reliance on trade has changed in the last few decades. To that end, we recon-
struct the global network of international food trade (in kilocalories of food traded/year) and investigate
its temporal dynamics. We also relate the food calorie network to the virtual water network [Carr et al.,
2012, 2013] and evaluate the WUE of the traded commodities.
TradedataweretakenfromtheFAOSTATdatabase[FAO STAT , 2012], which reports the trade ﬂow among
countries for a number of commodities between 1986 and 2010. In this study, a set of 153 countries (only
countries with population >1 million and for which production data were available) and 251 major food
commodities were analyzed (see Supporting Information, Table S1). Each commodity was associated with
a country-speciﬁc water footprint (m3/t)takenfromthestudybyMekonnen and Hoekstra,  and a
caloric content (kcal/t) taken from the FAO “Nutritive Factors” database [FAO STAT , 2012]. The trade of each
commodity was then converted into ﬂuxes of both virtual water (VW) and food calories by multiplying
the mass of the exchanged commodity by its water footprint and caloric content, respectively. The vir-
tual water network was then constructed considering each country or territory reported in the FAOSTAT
database as a network node connected to other nodes by links representing trade relationships. In the
VW network the strength of each (directed) link was calculated as the sum of the VW ﬂuxes associated
with all commodities traded along that link [e.g., Carr et al., 2012, 2013]. Likewise, the food calorie network
was determined by calculating for each link the total calories traded as the sum for all the commodities
exchanged through that link.
Production rates were available for each crop type from the FAOSTAT database [FA OS TAT , 2012]. For each
country, the total amount of water used for food production was calculated by multiplying the water
footprint of each commodity by the corresponding yield, and then by adding up these values for all the
food commodities included in this study [Carr et al., 2012]. The same approach was used to calculate the
calories of the food commodities produced by each country. The analysis of food calorie production and
D’ODORICO ET AL. © 2014 The Authors. 2
of the associated consumption of virtual water is aﬀected by problems of double accounting that were
addressed as follows.
In the case of secondary products —which are products derived from primary crops (e.g., bread is a
derived product of wheat; meat and milk can be derived products of feed)— caution was used to avoid
double accounting of the water resources required for their production. If a country produces both wheat
and bread, and part of the wheat is used to make the bread, by adding the water footprint of both we
would count part of the water consumption twice. Likewise, a similar problem of double accounting can
emerge from the calculation of the calories of food produced by each country.
Because the water footprint of primary crops is much higher than the water consumption of processing,
all secondary products were not included in the calculation of the water footprint and caloric content of
food, except for animal-based food products such as meat, milk, or eggs, which were treated as primary
products, while all the products derived from them were treated as secondary (Supporting Information,
Table S1). The animal-based products can be either from feed-fed animals or from livestock raised in
rangelands. While feed is included in the crop data reported by FAOSTAT, forage from rangeland and
pasture is not. Therefore, in this study we considered both a production, P, that included crops used as
feed and other uses, and a production, P′, obtained subtracting from Pthe amounts used as feed and
other nonfood uses; we included (both in Pand in P′) all the animal-based food products that are here
treated as primary (Supporting Information, Table S1) but we excluded all the secondary products. Thus,
while Pis the total food production (with no double accounting of primary and derived products), P′is
the food production for human consumption. The fraction of food production that is available for human
consumption can be calculated as 𝛽=P′/P(Supporting Information, Table S2). The feed (seed and “other
uses”) data were obtained from the Food Balance Sheets which were available only until the year 2009
[FAO STAT , 2012].
Because a similar issue of double accounting does not emerge in the case of trade, in the analysis of the
global patterns of trade we did not remove the secondary products. If a country produces and exports
bread made with wheat imported from another country (or produced domestically), the net calorie and
virtual water export would be calculated as the diﬀerence between bread exports and wheat imports
(or the whole bread exports if domestic wheat was used) without incurring into any double accounting.
Therefore, while secondary products were not removed from the trade data, we considered two types
of trade, namely, the total trade Tand the portion of trade for direct human consumption, T′,whichwas
calculated as T′=𝛽T. It has been noted that food production and trade can be analyzed and interpreted
in more than one way [Giampietro and Mayumi, 2000]. While T′is the total food trade for human con-
sumption, from a diﬀerent perspective global food security depends also on the trade of feed, which is
included in T. The water use eﬃciencies of production (P′)andtrade(T′) were then calculated for each
country as the ratio between that country’s total calories of production (P′)ortrade(T′) divided by the
corresponding water footprint. The calculation of country-speciﬁc water use eﬃciencies should have used
crop water requirements for the country of provenance of each commodity. However, it is not possible
to bring the calculations to this level of accuracy because the available data do not provide information
on the amounts of imported goods used for food, feed or other uses. To develop a consistent trade net-
work throughout the study period (1986– 2010), countries that have changed in recent years (e.g., the
uniﬁcation of Germany or Yemen, the splitting of Czechoslovakia or the disintegration of Yugoslavia and
the Soviet Union) were dealt with by considering their more aggregated (i.e., uniﬁed) conﬁguration, as
described by Carr et al. .
Globally, 80% of the human diet (measured in terms of caloric content) is accounted for by the 13 prod-
ucts reported in Table 1. More than 50% of the global diet is contributed by wheat (20%), rice (16%),
maize (13%), and soybean (8%), which account for 61% of the global production of protein for human
intake. These values refer to 2009 and include only food for direct human consumption (i.e., excluding
feed or crops for nonfood uses). The analysis of global trade data shows that 80% of the food calories
that are traded are contributed by 15 products, which account for 52% of the protein trade for human
consumption (Table 2). In 2009 more than 50% of the global food trade (in caloric content) was accounted
D’ODORICO ET AL. © 2014 The Authors. 3
Tab le 1 . The Major Food Commodities That Explain 80% of the Total Food Calories Produceda
Food Cal Protein Food Cal Protein
Commodity % Cal Cum. % Prot. Cum 𝛽WUE Commodity % Cal Cum. % Prot Cum. 𝛽WUE
Wheat 23.7 23.7 19.7 19.7 0.78 1.321 Wheat 20.4 20.4 17.2 17.2 0.81 1.321
Rice (milled equivalent) 18.4 42.1 11.6 31.3 0.95 0.970 Rice (milled equivalent) 16.1 36.5 10.4 27.6 0.88 0.970
Maize 9.7 51.9 10.7 42.0 0.33 1.298 Maize 12.8 49.3 12.7 40.2 0.40 1.298
Soy beans 5.3 57.2 14.2 56.3 0.97 0.356 Soybean 8.0 57.2 20.5 60.8 0.96 0.356
Sugarcane 4.7 61.9 0.7 57.0 0.97 0.752 Sugarcane 5.4 62.7 0.8 61.6 0.98 0.752
Pig meat 3.4 65.3 3.0 60.0 1.00 0.834 Pig meat 3.8 66.5 3.2 64.8 1.00 0.834
Sugar beet 3.3 68.6 1.48 61.5 0.96 2.465 Rape and mustard seed 3.1 69.6 2.9 67.8 0.91 1.515
Barley 2.6 71.3 5.52 67.0 0.26 1.342 Potatoes 2.1 71.8 1.1 68.8 0.86 1.782
Potatoes 2.5 73.7 1.24 68.3 0.74 1.782 Barley 2.1 73.8 2.5 71.3 0.38 1.342
Sorghum 1.9 75.6 2.1 70.4 0.46 0.674 Poultry meat 1.9 75.7 2.5 73.8 0.99 0.616
Rape and mustard seed 1.6 77.2 1.5 71.9 0.93 1.515 Vegetables, other 1.8 77.5 0.8 74.7 0.94 0.465
Groundnuts (shelled eq) 1.4 78.7 1.5 73.5 0.99 0.918 Sugar beet 1.7 79.2 0.7 75.4 0.94 2.465
Bovine meat 1.4 80.0 3.7 77.3 1.00 0.141 Groundnuts (shelled eq) 1.5 80.7 1.6 77.0 0.96 0.918
aFor each of these commodities we also report their contribution to the total production of protein for human consumption (expressed as a % of the total)
and the cumulated contribution. These values are based on production data (P′) for direct human consumption (i.e., excluding secondary products, feed
and other nonfood uses); 𝛽is the fraction of food production that is available for human consumption. Water use eﬃciency (WUE) is expressed in kcal/L.
for by wheat (22%), soybean (13%), palm oil (8%), and maize (7%). Interestingly, there have been no major
changes in the main food products traded for direct human consumption between 1986 and 2009, except
for an increase in the trade of palm oil. The amount of food calories shown in Figure 1 is not trivial when
compared with other major rates of energy consumption and trade sustaining the “metabolism” of our
societies: globally, the caloric content of food trade is about 13% of the global crude oil trade, while the
caloric content of all food products for direct human consumption is about 75% of the global consump-
tion of electric energy, and 10% of the total energy consumption— including energy used for power
generation, heating, transportation, and industrial uses, both from fossil fuels and renewable sources
(Supporting Information, Table S3).
There are some regional diﬀerences in the major food commodities (Supporting Information, Table S4):
more than 50% of the food calories produced in 2009 were from rice (32%) and wheat (22%) in Asia; soy-
bean (33%) and sugarcane (24%) in South America; maize (33%) and sugarcane (22%) in Central America;
maize (41%) and soybean (19%) in North America; wheat (36%) and barley (10%) in Europe; and wheat
(50%) and barley (14%) in Oceania. While the major food commodities produced in Asia, Africa, Europe,
Central America, and Oceania remained overall the same between 1986 and 2009, important changes
occurred in South America, where the production of soybean has increased, concurrently with an increase
in soybean exports from Brazil and Argentina to China.
The total amount of food produced (kcal/y) has steadily increased over the past three decades (Figure 1a).
Likewise, the total amount of food traded has also increased (Figure 1b). The increase in production is not
only due to the rise in food demand associated with demographic growth. Globally, the amount of food
produced per capita (Figure 1c) has also increased, which suggests the occurrence of some important
dietary changes with an average increase in calorie uptake per capita and, possibly, an increasing trend in
food waste. Interestingly, over the past three decades the fraction of food for direct human consumption
that is traded in the international market has increased from 15% in 1986 to 23% in 2009 (Figure 1d). Such
an increase is contributed by food from both plant and animal sources. Thus, presently, about one fourth
of food production is traded, implying that roughly one fourth of our current reliance on food is through
D’ODORICO ET AL. © 2014 The Authors. 4
Tab le 2 . The Major Food Commodities That Explain 80% of the Total Food Calories Tradeda
Food Cal Protein Food Cal Protein
Commodity % Cal Cum. % Prot. Cum. 𝛽WUE Commodity % Cal Cum. % Prot. Cum. 𝛽WUE
Wheat 28.5 28.5 24.4 24.4 0.78 1.321 Wheat 22.4 22.4 16.1 16.1 0.78 1.321
Soybean 10.2 38.6 21.8 46.2 0.97 0.356 Soybean 13.3 35.6 27.6 43.7 0.97 0.356
Sugar (raw equivalent) 9.1 47.7 0 46.2 0.99 1.970 Palm oil 7.9 43.5 0 43.7 0.61 1.878
Maize 8.0 55.8 9.9 56.3 0.33 1.298 Maize 7.5 51.0 0 43.7 0.33 1.298
Rice (milled equivalent) 4.4 60.2 0 56.3 0.95 0.970 Sugar (raw equivalent) 7.2 58.2 0 43.7 0.99 1.970
Palm oil 3.8 63.9 0 56.3 0.61 1.878 Rape and mustard seed 4.3 62.6 3 46.9 0.93 1.515
Soybean oil 2.8 66.8 0 56.3 0.92 0.480 Rice (milled equivalent) 4.1 66.6 0.1 47.0 0.95 0.970
Barley 2.3 69.1 6.9 63.4 0.26 1.342 Soybean oil 3.0 69.6 0 47.0 0.92 0.480
Rape and mustard seed 2.0 71.1 1.6 65.0 0.93 1.515 Pig meat 2.2 71.8 0.6 47.6 1.00 0.834
Sunﬂower seed oil 2.0 73.1 0 65.0 0.94 0.933 Sunﬂower seed oil 2.2 74.0 0.0 47.6 0.94 0.933
Cassava 1.9 75.0 0 65.0 0.59 1.081 Barley 2.1 76.1 2.8 50.4 0.26 1.342
Fats, animals, raw 1.8 76.7 0 65.0 0.45 2.195 Cocoa beans 1.4 77.6 0.1 50.5 0.99 0.081
Sorghum 1.5 78.3 1.8 66.8 0.46 0.674 Oilcrops, other 1.2 78.8 0 50.5 0.94 0.387
Oilcrops, other 1.4 79.7 0 66.8 0.94 0.387 Rape and mustard oil 1.2 79.9 0 50.5 0.87 1.461
Rape and mustard oil 1.4 81.0 0 66.8 0.87 1.461 Poultry meat 1.1 81.0 1.2 51.7 0.99 0.617
aFor each of these commodities we also report their contribution to the total trade of protein for human consumption (expressed as a % of the total) and
the cumulated contribution. These values are based on trade data (T′) for direct human consumption (i.e., excluding secondary products, feed and other
nonfood uses); 𝛽is the fraction of food production that is available for human consumption. Water use eﬃciency (WUE) is expressed in kcal/L.
This global scale analysis hides some important regional diﬀerences. From 1986 to 2009, food production
has increased mostly in South America (≈121%), Africa (≈81%), Asia (≈58% increase), and North Amer-
ica (≈57%), while yields in Europe have been stagnating (Figure 2). This pattern is in agreement with a
global assessment of the yield gap (i.e., the diﬀerence between actual and maximum potential crop yields)
based on agricultural production data for the year 2000 [Mueller et al., 2012]: while Western Europe had a
small potential for crop yield increase, Africa, South and Southeast Asia, South America, and— to a lesser
extent—North America still had relatively large yield gaps.
To place these numbers in the perspective of global food security we calculated the number of people
who could be fed by the food that is produced and traded. To this end, a balanced diet of 2700 kcal/d
(including waste [Porkka et al., 2013]) was used as a reference value [e.g., Falkenmark and Rockstrom, 2006;
Cassidy et al., 2013]. By considering this average caloric content, this analysis does not account for the
composition of the diet (e.g., the proportion of carbohydrates, proteins, and fats), nor for the geographic
diﬀerences in the current rates of food consumption. Nevertheless, these baseline calculations show
how, on average, the food production globally would be suﬃcient to feed more than 9.4 billion people
(Figures 1c and 3). This estimate is overall consistent with a recent study by the FAO that showed how
the current rates of food production would be suﬃcient to feed about 12 billion people with a lower
baseline diet of 2400 kcal/d per capita [FAOS TAT , 2012]. Likewise, the number of people who could be
fed by the traded food accounts for about 2 billion people (Figure 3). Thus, with the current rates of food
production (and without accounting for food waste), the global food production should be suﬃcient to
meet the demand of the world’s population (Figure 3). However, even though access to food is recognized
as a human right [U.N., 1948], today more than 10% of the global population is still undernourished [FAO,
When performed at the country scale, this analysis allows us to relate food production to the number of
people it could sustain, and to compare this number to the actual population of that country. Figure 4
shows a global map of caloric self-suﬃciency, based on a diet of 2700 kcal/d per capita. As noted, this
D’ODORICO ET AL. © 2014 The Authors. 5
Figure 1. (a) Total amount of food traded (kcal/y); (b) global food production (kcal/y); (c) global food production per capita
(kcal/y/cap); (d) percentage of global food production that is traded (%). Thin lines refer to food production and trade for direct
human consumption (P′and T′). Thick lines refer to food production, P, including feed and other uses (but excluding secondary
products) and food trade, T, including secondary products, feed and other uses. In panel (d) the thin and thick lines coincide.
Figure 2. Global food production by regions. Based only on P′, i.e., primary
plant and animal products available as food for direct human consumption (i.e.,
excluding feed and other uses).
analysis, however, does not account
for the impact of diet composition on
food self-suﬃciency. Protein deﬁciency
may lead to physiological unbalance,
particularly in young children, and is a
major cause of malnutrition in sub-
sistence societies, whose diets are
often poor in proteins and relatively
rich in carbohydrates [e.g., Cuny, 1999].
Examining the diet in terms of total
caloric intake, however, is an impor-
tant ﬁrst step in the assessment of
food security. The eﬀect of trade on
food security is shown in Figure 4:
diﬀerent shadings are used to indi-
cate whether the per capita calorie
requirement of the reference diet
(2700 kcal/d) can be met considering
food commodities available through
domestic production and trade. We
denote this condition as “suﬃciency.”
D’ODORICO ET AL. © 2014 The Authors. 6
Figure 3. Number of people who could be fed by the global food production
with an average balanced diet of 2700 kcal/d (thick line); global population
(dashed line) and number of people who could be fed by the food traded in the
international market (thin solid line). Based on production and trade data (P′
and T′) for direct human consumption (i.e., excluding secondary products, feed
and other nonfood uses).
North Africa and the Middle East are
overall not self-suﬃcient but rely on
trade to meet their food requirements
(Figure 4). Conversely, most of the
Sahel and East Africa are not suﬃ-
cient (Figure 4) despite their net food
imports (Figure 5). The United States,
Canada, Argentina, Brazil, Indonesia,
Malaysia, France, and Australia are
both major food producers [Porkka
et al., 2013] (Supporting Informa-
tion, Figure S1) and major exporters
(Table 3); despite their substantial
exports, these countries maintain their
suﬃciency state. India, Pakistan, and
China are also major food producers
(in absolute terms, Supporting Infor-
mation, Figure S2). Despite its net
imports (Figure 5), India has remained
“barely suﬃcient” throughout the
study period (Figure 4). Conversely,
Pakistan is not self-suﬃcient and
relies on trade to maintain access to
a barely suﬃcient amount of food
(Figure 4). The case of China is diﬀerent because it is a net importer despite its self-suﬃciency (Figure 4)
and its excess in food availability (Supporting Information, Figure S1). Similarly, most of Western Europe
is self-suﬃcient and a net importer. On the other hand, in a number of African countries as well as in
Afghanistan and Mongolia food imports are not strong enough to provide food suﬃciency (Figures 4
and 5). Trade does not seem to induce a loss in self-suﬃciency (except for Kazakhstan in 2000), nor an
increase in water stress in countries that are not self-suﬃcient. At most, we have found that a limited
number of self-suﬃcient countries [Zambia (2000), Bolivia (2000), and Indonesia (2010)] became barely
self-suﬃcient as an eﬀect of trade (Figure 4). This suggests that trade does not seem to erode the global
The increasing globalization of food in the course of the past three decades is reﬂected not only in the
rising volume of traded food calories, but also in the increase in the interconnection of the global food
calorie network (Figure 5). The total number of trade links in the network has increased from 8004 in 1986
to 13,945 in 2009 while the average number of links per country has more than doubled (Figure 5c). There
have also been some changes in the major contributors to net food exports (Table 3): while in 1986 the
United States contributed to 26% of the food placed on the global trade market (evaluated in terms of net
trade ﬂows), by 2009 it had decreased to 17%. This decline in the predominance of the United States as a
net food exporter was paralleled by a similar decline of France and the emergence of Indonesia and Brazil
as major exporters. Interestingly, 50% of the net exports are controlled by about ﬁve countries (Table 3).
It is also interesting to relate the availability and trade of food calories to the resources used in the produc-
tion process. We focus in particular on water, which remains a major limiting factor for agriculture. Food
production requires more water than any other human activity; it has been estimated that most of the
human appropriation of freshwater resources (80%– 90%; e.g., Falkenmark et al., 2004) is used in agricul-
ture. Some countries do not have enough renewable water resources to produce all the food they need
and therefore depend on imports from other regions. Thus, trade is associated with a virtual transfer of
the water used in the production of the traded commodities (Allan, 1998). Recent research has investi-
gated the global patterns of virtual water trade (Hoekstra and Chapagain, 2008; Carr et al., 2012), while
their relationship with food production and trade remains poorly understood. We found that the WUE of
the traded commodities was about the same as that of production in 1986 and it declined in the course
of the study period (Table 4). This suggests that in the last few decades the increase in food trade has
D’ODORICO ET AL. © 2014 The Authors. 7
Figure 4. Global maps of country suﬃciency and self-suﬃciency deﬁned by comparing the per capita food production to a diet of 2700 ±10% kcal/person/d (including waste).
Countries were classiﬁed as not self-suﬃcient if production <2430 kcal/d; barely self-suﬃcient if the production was between 2430 and 2970 kcal/d; self-suﬃcient if the
production >2970 kcal/d. Likewise, countries were classiﬁed as not suﬃcient, barely suﬃcient, and suﬃcient if their average food supply per capita (accounting both for
production and net trade) was below, in between, or above those two values, respectively.
focused on food commodities with relatively low WUE. To better understand these patterns we have
considered plant, animal, and luxury products separately. The WUE of animal-based food products is
overall much lower than that of plant products (about 30%). The WUE of the traded plant products has
decreased over time. A (smaller) decline is observed also for the traded animal products, while the luxury
products have maintained an almost constant WUE. The traded animal products have a higher WUE than
the total production of animal-based food commodities, while the opposite is true about luxury products
(Table 4). In the case of plant products, at the beginning of the study period the WUE of trade was higher
than that of production, but it then decreased below the WUE of production. The global distribution of the
WUE of food production does not reﬂect any obvious climatic pattern, while it appears to overall increase
with countries’ aﬄuence (Figure 6).
Decades of research on social development and famine indicate that food security depends not only
on food availability and the ability of the planet to produce enough to nourish the entire global
population [Malthus, 1798; Cohen, 1995; Lee, 2011], but also on other factors that aﬀect the access to food,
including trade [Sen, 1981], political stability and lack of conﬂict [Deveroux, 2001], and the existence of
institutions and norms of resource governance [e.g., Ostrom, 1990]. By evaluating countries’ reliance on
national production and international trade, this study developed a country-scale analysis of food security
with a focus on food suﬃciency and availability. Several countries around the world are not in conditions
of food self-suﬃciency but depend on imports from other regions [Fader et al., 2013; Suweis et al., 2013].
In the past two to three decades, most of Africa and the Middle East have not been self-suﬃcient. Trade,
however, appears to have improved the situation in the Sahel region, which is now closer to conditions
of suﬃciency than in the previous decades (Figure 4). In other countries of sub-Saharan Africa and central
Asia, however, trade was unable to eradicate food insuﬃciency. At the same time, only few countries
lost their self-suﬃciency because of trade. Thus, trade seems to be enhancing rather than eroding food
security. Overall, in the last two decades there has been an increase in the number of (trade-dependent)
countries that reach suﬃciency through their reliance on trade. The ongoing intensiﬁcation of trade
and the fact that, globally, about 23% of the food is traded suggest that global food security can be
D’ODORICO ET AL. © 2014 The Authors. 8
Figure 5. The global network of food trade in 1986 (a) and 2010 (b). (c) Changes in the average number of export links per country
(or “degree”) during 1986–2010.
threatened not only by regional climate extremes (drought, ﬂood, frost) but also by price volatility and
changes in the food market [Headley, 2010]. Countries that strongly rely on trade are expected to be par-
ticularly vulnerable, especially if their economies are not strong enough to absorb the shocks of food price
volatility in the global market [FAO-OECD, 2011; Fader et al., 2013; D’Odorico et al., 2010]. At the same time
trade plays a crucial role in allowing societies in conditions of food deﬁcit to meet their demand through
imports from other regions of the world.
D’ODORICO ET AL. © 2014 The Authors. 9
Tab le 3 . Major Contributors to Net Food Exports
Country % % cum Country % % cum
United States 25.8 25.8 United States 17.0 17.0
France 8.9 34.7 Brazil 9.9 26.9
Canada 7.9 42.6 Argentina 8.5 35.4
Australia 6.7 49.3 Indonesia 5.9 41.3
Argentina 6.4 55.7 France 5.9 47.1
Thailand 3.8 59.4 Canada 5.6 52.8
Malaysia 3.7 63.1 Malaysia 5.4 58.1
Brazil 3.5 66.6 Germany 3.4 61.5
United Kingdom 3.0 69.6 Australia 3.1 64.6
China. mainland 2.9 72.4 Netherlands 3.1 67.7
Tab le 4 . Global Water Use Eﬃciency of Food Commodities Produced and Traded (kcal/L)
WUE Production Trade
(kcal/L) 1986 2010 1986 2010
All products 1.372 1.346 1.392 1.206
Plant 1.736 1.640 1.931 1.601
Animal 0.532 0.528 0.695 0.622
Luxury 1.443 1.528 0.486 0.480
Because of the globalization of food and the strong interconnection existing within the food trade net-
work, the eﬀect of episodic regional declines in crop production can be felt globally [Suweis et al., 2013].
For instance, the recent food crisis of 2007 was likely induced by droughts in Russia, Ukraine, and the
United States, and an increase in global crop demand for agrofuels. The consequently escalating food
prices caused riots and social unrest in some developing countries [Hermele, 2014]. To curb the food price
increase, some governments banned exports, while trade-dependent countries started to panic [Maetz
et al., 2011]. This food crisis and the episodes that followed in 2011 demonstrate that uncertain and unre-
liable food markets can lead to food insecurity [FAO-OECD, 2011; Headley, 2010; Welton, 2011]. This eﬀect
is expected to become even stronger as the reliance on trade and the globalization of food increase [Fader
et al., 2013; Suweis et al., 2013]. However, the ongoing development of tissue engineering technologies
for in vitro meat production might provide in a near future new means to enhance food security with-
out requiring soils and water, though the environmental and ethical implications of such transition in the
meat industry is still diﬃcult to foresee [Tuomisto and Teixeira de Mattos, 2011; Post , 2012].
To date water remains one of the major factors constraining food production. The WUE of food production
tends to increase with aﬄuence (Figure 6b), which presumably reﬂects access to modern water eﬃcient
technologies in the economically more developed countries.
It has been observed that the international trade of food commodities induces a virtual transfer of embod-
ied land [Kitzes et al., 2009], water [Allan, 1998], carbon [Kastner et al., 2011], nitrogen [Galloway et al.,
2007], and other land based resources [e.g., Hermele, 2014], while most of the environmental impacts
of agricultural production remain in the producing countries [Meyfroidt et al., 2013]. Thus, trade is asso-
ciated with the globalization of resources [e.g., Hoekstra and Chapagain, 2008], the externalization of
environmental impacts [e.g., O’Bannon et al., 2014], and the establishment of teleconnections between
those impacts and their drivers (e.g., consumer behavior, price volatility, or climate ﬂuctuations) [DeFries
et al., 2010; Schmitz et al., 2012; Meyfroidt et al., 2013]. An often overlooked aspect of international trade is
D’ODORICO ET AL. © 2014 The Authors. 10
Figure 6. Global map of WUE (kcal/L) of food production for 2010 (based on all
food commodities). (a) Global map; (b) dependency on per capita gross
domestic product (GDP). Averages (squares) and standard deviations (error bars)
were calculated after binning the raw data.
its impact on food security [Fader
et al., 2013]. This study reconstructed
the global network of food calorie
trade and evaluated its changes in the
course of the past three decades. We
found that, globally, about one fourth
of the food produced for human
consumption is traded internation-
ally. While the caloric content of the
food produced worldwide would be
suﬃcient to feed the whole global
population, there are countries in
conditions of chronic food scarcity.
In these countries the food demand
by the local population by far exceeds
the supply allowed by the land, water,
climate, and soils locally available for
food production (Figure 4 and Sup-
porting Information, Figure S2). This
unbalanced condition can be sus-
tained either by accepting that actual
food consumption does not meet
the demand (hence the persistence
of malnourishment) or by relying on
international trade for the redistri-
bution of food commodities among
diﬀerent regions of the world.
Allan, J. A. (1998), Virtual water: A strategic resource global solutions to regional deﬁcits, Ground Water,36(4), 545–546.
Boserup, E. (1981), Population and Technological Change, University of Chicago Press, Chicago, Ill.
Brown, L. R. (2013), Full Planet, Empty Plates: The New Geopolitics of Food Scarcity, pp. 161, W.W. Norton & Company Publishing,
Carr, J. A., P. D’Odorico, F. Laio, and L. Ridolﬁ (2012), On the temporal variability of the virtual water network, Geophys. Res. Lett.,39,
Carr, J. A., P. D’Odorico, F. Laio, and L. Ridolﬁ (2013), Recent history and geography of virtual water trade, PLOS One,8(2), e55825,
Cassidy, E. S., P. C. West, J. S. Gerber, and J. A. Foley (2013), Redeﬁning agricultural yields: From tonnes to people nourished per
hectare, Environ. Res. Lett.,8, 034015.
Cohen, J. E. (1995), How Many People Can the Earth Support?, WW Norton & Company, New York.
Coll, M., S. Libralato, S. Tudela, I. Palomera, and F. Pranovi (2008), Ecosystem overﬁshing in the ocean, PLoS One,3(12), e3881,
Cuny, F. C. (1999), Famine, Conﬂict and Response, Kumarian Press, West Har tford, Conn.
DeFries, R. S., T. K. Rudel, M. Uriarte, and M. Hansen (2010), Deforestation driven by urban population growth and agricultural trade in
the twenty-ﬁrst century, Nat. Geosci.,3, 178 –181.
Delgado, C. L. (2003), Rising consumption of meat and milk in developing countries has created a new food revolution, Am. Soc. Nutr.
Sci.,133(Suppl. 2), 3907– 3910.
Deveroux, S. (2001), Sen’s entitlement approach: Critiques and counter-critiques, Oxford Dev. Stud.,29(3), 245– 263.
D’Odorico, P., and M. C. Rulli (2013), The fourth food revolution, Nat. Geosci.,6(6), 417 –418, doi:10.1038/ngeo1842.
D’Odorico, P., F. Laio, and L. Ridolﬁ (2010), Does globalization of water reduce societal resilience to drought?, Geophys. Res. Lett.,37,
Duarte, C. M., et al. (2009), Will the oceans help feed humanity?, BioScience,59, 967– 976.
EISA (2007), Energy Independence and Security Act of 2007, Public Law 110– 140, 110th Congress of the United States of America,
Washi ngton, D. C.
Erb, K. H., F. Krausmann, W. Lucht, and H. Haberl (2009), Embodied HANPP mapping the spatial disconnect between global biomass
production and consumption, Ecol. Econ.,69, 328– 334.
Data used in this manuscript are pub-
licly available online through the FAO-
STAT data base (http://faostat.fao.org).
D’ODORICO ET AL. © 2014 The Authors. 11
Earth’s Future 10.1002/2014EF000250
Erisman, J. W., M. A. Sutton, J. Galloway, Z. Klimont, and W. Winiwarter (2008), How a century of ammonia synthesis changed the
world, Nat. Geosci.,1, 636 –639.
E.U. (2009), Directive 2009/28/EC of the European Parliament and of the Council on the promotion of the use of energy from
renewable sources and amending and subsequently repealing Directives 2001/77/EC and 2003/30/EC., Oﬃcial Journal of the
European Union L 140, 5 June 2009, pp. 16– 47.
Fader, M., D. Gerten, M. Krause, W. Lucht, and W. Cramer (2013), Spatial decoupling of agricultural production and consumption:
Quantifying dependences of countries on food imports due to domestic land and water constraints, Environ Res. Lett.,8, 014046.
Falkenmark, M., and J. Rockstrom (2006), The new blue and green water paradigm: Breaking new ground for water resources
planning and management, J. Water Resour. Plann. Manage.,132(3), 129–132.
Falkenmark, M. J., J. Rockström, and H. Savenjie (2004), Balancing Water for Humans and Nature, Earthscan, London, U. K.
FAO (2012), Statistical Yearbook 2012 – World Food and Agriculture. Part 3: Feeding the World, FAO, Rome, Italy.
FAO-OECD (2011), Food and Agriculture Organization/Organization for Economic Co-operation and Development: Price Volatility in Food
and Agricultural Markets: Policy Responses. Food and Agriculture Organization of the United Nations, Rome, Italy.
FAOSTAT (2012), Food and Agriculture Organization (FAO): FAOSTAT. [Available at http://faostat.fao.org].
Foley , J. A., et al. (2011), Solutions for a cultivated planet, Nature,478, 337 –342.
Galloway, J. N., et al. (2007), International trade in meat: The tip of the pork chop, Ambio,36(8), 622–629.
Gephart, J., M. L. Pace, and P. D’Odorico (2014), Freshwater savings from marine protein consumption, Environ. Res. Lett.,9, 014005,
Giampietro, M., and K. Mayumi (2000), Multiple-scale integrated assessment of societal metabolism: Introducing the approach, Popul.
Environ.,22(2), 109– 153.
Godfray, H. C. J., et al. (2010), Food security: The challenge of feeding 9 billion people, Science,327, 812–818.
Headley, D. (2010), Rethinking the global food crisis: The role of trade shocks, Food Policy,36, 136– 146.
Hermele, K. (2014), The Appropriation of Ecological Space, Routledge, New York.
Hoekstra, A., and A. Chapagain (2008), Globalization of Water, Malden, Mass., Wiley Blackwell, Hoboken, N. J.
IPCC (2013), Summary for policymakers, in Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth
Assessment Report of the Intergovernmental Panel on Climate Change,editedbyT.F.Stocker,D.Qin,G.-K.Plattner,M.Tignor,S.K.
Allen, J. Boschung, A. Nauels, Y. Xia, V. Bex, and P. M. Midgley, Cambridge Univ. Press, Cambridge, U.K.
Kastner, T., K.-H. Erb, and S. Nonhebel (2011), International wood trade and forest change: A global analysis, Global Environ. Change,
21, 947– 956.
Kitzes, J., et al. (2009), A research agenda for improving national Ecological Footprint accounts, Ecol. Econ.,68, 1991– 2007.
Lee, R. (2011), The outlook for population growth, Science,333, 569– 573.
MacDonald, G. K. (2013), Eating on an interconnected planet, Environ. Res. Lett.,8, 021002, doi:10.1088/1748-9326/8/2/021002.
Maetz M, Aguirre M, Kim S, Matinroshan Y, Pangrazio G, Pernechele V. 2011. Food and agricultural policy trends after the 2008 food
security crisis. Renewed attention to agricultural development. Resources for Policy Making. Applied Work EASYPol Module 125,
FAO, Rome, Italy.
Malthus, T. (1798), in An Essay on the Principle of Population, edited by A. Flew, 1st ed., pp., Penguin English Library, London, U. K.
Mekonnen, M. M., and A. Y. Hoekstra (2012), A global assessment of the water footprint of farm animal products, Ecosystems,15,
Meyfroidt, P., E. F. Lambin, K.-H. Erb, and T. W. Hertel (2013), Globalization of land use: Distant drivers of land change and geographic
displacement of land use, Curr. Opin. Environ. Sustain,5, 438– 444.
Mueller, N. D., et al. (2012), Closing yield gaps through nutrient and water management, Nature,490, 254– 257.
O’Bannon, C., J. A. Carr, D. A. Seekell, and P. D’Odorico (2014), Globalization of agricultural pollution due to international trade, Hydrol.
Earth Syst. Sci.,18, 503 –510, doi:10.5194/hess-18-503-2014.
Ostrom, E. (1990), A general framework for analyzing sustainability of social-ecological systems, Science,325(5939), 419–422.
Pimentel, D., L. E. Hurd, A. C. Bellotti, M. J. Forster, I. N. Oka, O. D. Sholes, and R. J. Whitman (1973), Food production and the energy
crisis, Science,182, 443–449.
Porkka, M., M. Kummu, S. Siebert, and O. Varis (2013), From food insuﬃciency towards trade dependency: A historical analysis of
global food availability, PLoS One,8(12), e82714, doi:10.1371/journal.pone.0082714.
Post, M. J. (2012), Cultured meat from stem cells: Challenges and prospects, Meat Sci.,92, 297–301.
Ray, D. K., N. Ramankutty, N. D. Mueller, P. C. West, and J. A. Foley (2012), Recent patterns of crop yield growth and stagnation, Nat.
Comm.,3, 1293, doi:10.1038/ncomms2296.
Schmitz, C., A. Biewald, H. Lotze-Campen, A. Popp, J. P. Dietrich, B. Bodirsky, M. Krause, and I. Weindl (2012), Trading more food:
Implications for land use, greenhouse gas emissions, and the food system, Global Environ. Change,22, 189–209.
Sen, A. (1981), Poverty and Famines: An Essay on Entitlements and Deprivation, Oxford Univ. Press, New York.
Suweis, S., A. Rinaldo, A. Maritan, and P. D’Odorico (2013), Water-controlled wealth of nations, Proc. Natl. Acad. Sci. U.S.A.,110(11),
Tilman, D., C. Balzer, J. Hill, and B. L. Befort (2011), Global food demand and the sustainable intensiﬁcation of agriculture, Proc. Natl.
Acad. Sci. U.S.A.,108, 20260– 20264.
Tuomisto, H. L., and M. J. Teixeira de Mattos (2011), Environmental impacts of cultured meat production, Environ. Sci. Technol.,
2011(45), 6117– 6123.
U.K. Oﬃce for Science (2011), Foresight. The future of food and farming, Final Project Report, The Government Oﬃce for Science,
London, U. K.
U.N. (1948), Universal declaration of human rights (UDHR), Article 25, United Nations.
U.N. Population Division (2012), World Population Prospects: The 2012 Revision, Department of Economic and Social Aﬀairs of the
United Nations Secretariat, New York.
Weinzettel, J., E. G. Hertwich, G. P. Peters, K. Steen-Olsen, and A. Galli (2013), Aﬄuence drives the global displacement of land use,
Global Environ. Change,23, 433– 438.
Welton, G. (2011), The impact of Russia’s 2010 grain export ban, Oxfam Research Reports,Oxford,U.K.
D’ODORICO ET AL. © 2014 The Authors. 12