Energy and food security implications of transitioning synthetic nitrogen fertilizers to net-zero emissions

For all countries, i, we estimated the number of people fed from synthetic nitrogen fertilizers, ${P_i}$, by considering country-specific agricultural use of nitrogen fertilizers, ${N_i}$ (kg N), daily nitrogen intakes in diets, $\mu$ (kg N/person), the fraction of nitrogen lost from farm to fork, $\delta$, and country-specific nitrogen use efficiency, ${\eta _i}$ (or the fraction of nitrogen lost on the field and that is not used to produced food):

$$\begin{equation}{P_i} = \frac{{{N_i}}}{{{\mu _i}}}\,\delta \,{\eta _i}\,.\end{equation} \tag{ 1 }$$

Country-specific nitrogen fertilizers usage are taken from the FAO database for year 2019 [26]. Country-specific daily nitrogen intakes in diets are derived from per capita daily protein intakes as provided by FAO for year 2019 [26] and considering that 18% of protein content is nitrogen [27]. Nitrogen is the building block of amino acids that are required to produce proteins. The FAO database reports country-specific annual statistics of dietary protein consumption with a global median protein intake of 84 g of protein per day per person; a minimum intake of 26 g of protein per day per person in the Democratic Republic of Congo and a maximum intake of 144 g of protein per day per person in Iceland [26].

Our food systems are highly inefficient, with most of the food being lost or wasted [28]. The share of nitrogen lost from farm to fork is assumed to be 42.5%, an average value derived from previous global estimates of 41% and 44% [3, 29]. Such estimates quantified production and losses from farm to fork of digestible protein in the global food system, accounting for: harvesting losses, post-harvest losses, non-food uses, processing and packaging losses, distribution waste, and consumption waste [3, 29]. We acknowledge that a more appropriate analysis would use up-to-date country-specific and crop-specific values of nitrogen lost from farm to fork. However, given the complexity of our food systems, such data are only partially available—see for example the analysis of Gustavsson et al at regional level [30]—and are affected by large uncertainties [31, 32]. This highlights the need for more accurate data of nitrogen wasted in food systems.

Nitrogen use efficiency represents the quantity of applied nitrogen incorporated in crops and is assessed as the ratio between crop nitrogen uptake and applied nitrogen fertilizer [5, 33]. It is a measure of how efficiently nitrogen fertilizers are used in agriculture. The global average nitrogen use efficiency is estimated to be 46%, indicating substantial losses (more than half of the total nitrogen used) of reactive nitrogen to the environment [5]. We use country-specific and publicly available nitrogen use efficiency statistics [33]. Importantly, the ongoing global increase in animal protein intakes is leading to a decline in overall nitrogen use efficiency [5].

We analyzed country-specific trade balances to quantify synthetic nitrogen fertilizers self-sufficiency and determine net-importers and net-exporters. Specifically, for all countries, i, we quantify synthetic nitrogen fertilizers produced domestically with and without embedded natural gas imports. When not accounting for natural gas imports, we assess the quantity of synthetic nitrogen fertilizers imported or exported, ${D_i}$ (kg N), by taking the difference between total domestic production of synthetic nitrogen fertilizer, ${T_i}$ (kg N), and domestic agricultural use of synthetic nitrogen fertilizer, ${N_i}$ (kg N):

$$\begin{equation}{D_i} = {T_i} - \,{N_i}\,\end{equation} \tag{ 2 }$$

${D_i}$ represents the trade deficit or surplus of synthetic nitrogen fertilizers in a specific country. When ${D_i}$ is smaller than zero, a country is a net-importer of synthetic nitrogen fertilizers; when ${D_i}$ is greater than zero, a country is a net-exporter.

When accounting for natural gas imports, hence for the coupling of energy and food systems, we quantify the quantity of synthetic nitrogen fertilizers produced domestically, ${D_{i,\mathrm{CH_4}}}$, by taking the difference between total production of synthetic nitrogen fertilizer made from domestic natural gas, ${T_{i,\,CH4}}$, and agricultural use of synthetic nitrogen fertilizer, ${N_i}$ (kg N):

$$\begin{equation}{D_{i,\mathrm{CH_4}}} = {T_{i,\mathrm{CH_4}}} - \,{N_i}.\end{equation} \tag{ 3 }$$

When ${D_{i,CH4}}$ is smaller than zero, a country is a net-importer of synthetic nitrogen fertilizers, at the net of imported natural gas. When ${D_{i,\mathrm{CH_4}}}$ is greater than zero, a country is a net-exporter. The discrepancy between ${D_i}$ and ${D_{i,\mathrm{CH_4}}}$ provides a measure of the exposure of the food system to energy trading and potential shocks in energy supply. The total production of synthetic nitrogen fertilizer made from domestic natural gas, ${T_{i,\mathrm{CH_4}}}$, is computed by accounting for the fraction of nitrogen fertilizers produced from imported natural gas:

$$\begin{equation}{T_{i,\mathrm{CH_4}}} = {T_i}\,\left( {\,1 - \frac{{{I_i}}}{{{C_i}}}} \right)\end{equation} \tag{ 4 }$$

where ${I_i}$ (m³) is the amount (volume) of imported natural gas, and ${C_i}$ (m³) is the overall natural gas consumption in each country (29). To the best of our knowledge, there are not available dataset that specify the purpose of natural gas import in each country. Therefore, we assume that the share of synthetic nitrogen fertilizers produced from imported natural gas is the same as the total share of imported natural gas in each country. Data of total production of synthetic nitrogen fertilizer, ${T_i}$ (kg N), and agricultural use of synthetic nitrogen fertilizer, ${N_i}$ (kg N) are taken from FAO for year 2019 [26]. The number of people fed from imported synthetic nitrogen fertilizers with and without natural gas imports (${D_i}$ and ${D_{i,\mathrm{CH_4}}}$) is assessed using equation (1) and considering country-specific daily nitrogen intakes in diets, $\mu$ _i (kg N/person), fraction of nitrogen lost from farm to fork, $\delta$, and country-specific nitrogen use efficiency, ${\eta _i}$.

Conventional production of ammonia (NH₃) is carried out via the Haber–Bosch process. Globally, conventional Haber–Bosch processes use natural gas (70%) and coal (26%) as feedstock, with oil and electricity accounting for less than 4% of global production [6]. The availability of feedstock and process energy is a key determinant of where and how NH₃ is produced. Low-cost natural gas in the United States, Middle East, and Russia explains the prominent role of these regions and their natural gas-based plant fleets. China's abundant coal reserves explain its heavy reliance on the fuel, which accounts for around 85% of its production. Here, we assume that all countries use natural gas, except for China (mix of 85% coal and 15% natural gas), United States, South Africa, and Indonesia (all featuring a mix of 20% coal and 80% natural gas) [6].

When based on natural gas, the process uses about 0.49 ton of fossil carbon (t C), or about 0.65 ton of natural gas, to produce one ton of ammonia (t NH₃); the coal-based process uses about 0.88 t C (or about 0.62 ton of coal) to produce one t NH₃ [6] (supplementary figure 2). By applying stoichiometry, business-as-usual NH₃ production with natural gas and coal emits 1.8 t CO₂/t NH₃ and 3.2 t CO₂/t NH₃, respectively (supplementary figure 2).

A stream of pure hydrogen needs to be fed into the Haber–Bosch reaction for NH₃ synthesis. Therefore, CO₂ is removed through primary and secondary steam methane reforming (SMR) and a two-stage water-gas shift (WGS). The concentrated stream of CO₂ can be used for urea-based fertilizers production or alternatively is vented to the atmosphere contributing to global warming. Similarly, the diluted CO₂ in the exhaust gases from fossil fuels combustion for energy inputs is emitted to the atmosphere. The Haber–Bosch process is composed of two SMR reactors in series. The first SMR reactor is endothermic, with the required heat being provided by external combustion of methane fuel. The second SMR reactor is autothermal, while the WGS reaction is exothermic, making a significant amount of heat available for heat integration. Such heat is typically used for producing high-pressure steam, which is expanded in steam turbines for compression purposes.

When adopting carbon capture and storage, the concentrated CO₂ resulting from hydrogen production, and the diluted CO₂ in the exhaust gases from fossil fuel combustion are captured, transported to a suitable storage site, and permanently sequestered [34]. Because CO₂ transport via pipeline results in negligible CO₂ emissions, we do not consider associated CO₂ emissions from CO₂ transport and storage [35].

For the business-as-usual route, we consider an average electricity consumption of 0.18 MWh t⁻¹ NH₃ when the process uses natural gas as feedstock (ranging from 0.08 MWh t⁻¹ NH₃ when the process adopts SMR, to 0.28 MWh t⁻¹ NH₃ when the process adopts auto-thermal reforming), and an average electricity consumption of 0.91 MWh t⁻¹ NH₃ for the coal-based process [6]. Such electricity consumption is additional to the energy inputs provided by fossil fuels and is used to run auxiliary equipment.

When implementing carbon capture and storage to concentrated and diluted CO₂ streams, the electricity consumption becomes 0.35 MWh t⁻¹ NH₃ (range 0.28–0.41 MWh t⁻¹ NH₃) for natural gas-based processes and 1.26 MWh t⁻¹ NH₃ for coal-based processes [6]. Overall, the CO₂ capture process has a capture efficiency of about 95% [6], with 5% of the emissions still escaping to the environment and being captured back with direct air capture to achieve net-zero CO₂ emissions. The capture efficiency of 95% is assumed as the optimal trade-off between higher CO₂ purity and higher capture costs [6]. This efficiency value, which is higher than typical values for post-combustion capture of about 90%, is made possible by the combination of the concentrated CO₂ resulting from hydrogen production (higher capture efficiency) and the diluted CO₂ in the exhaust gases (capture efficiency of about 90%) [6].

Various studies have assessed processes that go beyond fossil fuels in NH₃ production by using hydrogen produced from water electrolysis or biomass [36–42]. Both the electrification and the biomass routes still rely on the Haber–Bosch process to produce NH₃, which is either electrified or fueled by biomass.

The electrification route relies on the use of renewable energy to produce hydrogen and nitrogen for NH₃ synthesis [38, 39, 41]. In this case, electricity provides all the energy requirements, replacing fossil fuels as both feedstock and fuel. Hydrogen is produced via water electrolysis and converted to NH₃ through a Haber–Bosch process like the conventional one; nitrogen is produced through an air separation unit [6]. Such a process requires 0.18 t H₂ and 0.82 t N₂, with an average electricity consumption of 9.3 MWh t⁻¹ NH₃ (range 8.6–10 MWh t⁻¹ NH₃) [6, 43].

Biomass-based NH₃ production is carried out via a gasification process starting from dry wood chips, here assumed to have a 43% carbon content and a lower heating value of about 17 MJ kg⁻¹. The process uses about 2 ton of dry biomass and 0.37 MWh to produce one t NH₃ [36, 44]. This amount of biomass is used both as a fuel, to produce the heat required by the first reforming reactor, and as a feedstock. For all net-zero emissions routes, schematics of processes enabling net-zero emissions production of NH₃ are shown in supplementary figure 1.

The carbon intensity of all routes is computed by considering the carbon intensity of the production processes, as well as the carbon intensity (including the embedded emissions) of the technologies supplying electricity. Electricity can be supplied via: (i) electricity grid, with a carbon intensity varying country-by-country (world average equal to 458 kg of CO₂ equivalent per MWh of electricity) [45]; (ii) solar photovoltaic panels, which result in a median value of 87.5 kg CO₂eq per MWh of produced electricity (range 23–183 kg CO2eq MWh⁻¹) [46]; (iii) wind turbines, with a median carbon intensity of 11 kg CO2eq MWh⁻¹ (range 8–23 kg CO2eq MWh⁻¹) [46]; (iv) nuclear power plants, with a median carbon intensity of 12 kg CO2eq MWh⁻¹(range 4–110 kg CO2eq MWh⁻¹) [46]. Furthermore, we consider the lifecycle carbon emissions of the natural gas supply chain, which account for the emissions released upon use of natural gas (both as a fuel and as a feedstock), as well as for natural gas leaks along the supply chain; here we assume natural gas leaks equal to 1.5% of the required amount of natural gas needed to produce NH₃ under the business-as-usual and carbon capture and storage routes. Indeed, recent research demonstrated that natural gas leaks occur across the entire supply chain, including production, processing, pipeline transportation, and distribution [47–49]. Overall, natural gas leaks rate from 0.2% to 8%, with 1.5% being a reasonable likely value [50].

For all four routes $j$ and all countries $i$, the total CO₂ emissions, ${C_{i,j}}$, are obtained as:

$$\begin{equation}{C_{i,j}} = {A_i}\left( {{\epsilon _j} + {\alpha _j}\gamma } \right)\end{equation} \tag{ 5 }$$

where ${A_i}$ is the amount of produced NH₃ in country $i$, ${\epsilon _j}$ is the carbon intensity of the production process of route $j$, ${\alpha _j}$ is the specific electricity consumption of route $j$, and $\gamma$ is the carbon intensity of electricity generation (which depends on the considered generation technology among grid, solar, wind and nuclear). The emissions resulting from NH₃ production are non-zero only for the business-as-usual and the carbon capture and storage routes (for the latter, they include the 5% fugitive emissions from the capture process). However, for all net-zero routes the emissions resulting from electricity production must be offset via direct air capture, which also requires energy to operate [51]. The amount of required direct air capture to achieve net-zero emissions, ${D_{i,j}}$, is computed as:

$$\begin{equation}{D_{i,j}} = \left( {{C_{i,j}} + {C_{i,j}}\gamma \lambda } \right)\lambda \end{equation} \tag{ 6 }$$

where $\lambda$ is the amount of electricity required to capture one ton of CO₂ via direct air capture. This is the sum of electricity consumption and electrified heat consumption of direct air capture. For the calculations, we consider a direct air capture technology based on solid sorbents [51]. We consider an electricity consumption of 0.35 MWh t⁻¹ CO₂ (which already includes 0.1 MWh t⁻¹ CO₂ for CO₂ compression for making it ready for transport and storage) and a heat consumption of 1.75 MWh t⁻¹ CO₂ [51]. Heat is required at low temperature (around 100 °C) and can be supplied via heat pump (coefficient of performance of 4), resulting in $\lambda$ = 0.79 MWh t⁻¹ CO₂. The deployment of direct air capture with CO₂ storage can result in further environmental tradeoffs [52].

For all routes and for all countries, the total electricity consumption, ${E_{i,j}}$, is given by the contribution of NH₃ production and the one of carbon capture via direct air capture:

$$\begin{equation}{E_{i,j}} = {A_i}{\alpha _j} + {C_{i,j}}\gamma \lambda .\end{equation} \tag{ 7 }$$

The land use and water consumption of all routes are computed by considering on-site electricity demand for NH₃ production, hydrogen production, direct air capture, and biomass. The calculations are based on the data reported in supplementary table 2.

The total land use and water consumption are computed by multiplying the total electricity production, the amounts of required hydrogen, direct air capture, and biomass, by the corresponding factors in supplementary table 2. Accordingly, for all countries and routes, the land use, ${L_{i,j}}$, is given by:

$$\begin{equation}{L_{i,j}} = {E_{i,j}}a + {D_{i,j}}b + {A_i}{\beta _j}c\end{equation} \tag{ 8 }$$

where ${\beta _j}$ is the amount of biomass required per unit of NH₃ in route $j$ (only needed for the biomass route); $a$, $b$, and $c$, are the land use intensities of electricity generation, direct air capture, and biomass, respectively.

Similarly, for all countries and routes, the total water consumption, ${W_{i,j}}$, is given by:

$$\begin{equation}{W_{i,j}} = {E_{i,j}}d + {D_{i,j}}e + {A_i}\left( {{\beta _j}f + {\eta _j}g} \right)\end{equation} \tag{ 9 }$$

where ${\eta _j}$ is the amount of hydrogen required per unit of NH₃ in route $j$ (only needed for the electrification route); $d$, $e$, $f$, and $g$ are the water consumption intensities of electricity generation, direct air capture, biomass, and hydrogen production, respectively.

By considering country-specific nitrogen use efficiencies (i.e. the fraction of nitrogen lost on the field, which is not used to produced food) [33], nitrogen waste and losses in crops from farm to fork [3, 29], and country-specific per capita nitrogen intakes [26], we quantify the number of people that consume synthetic nitrogen fertilizers and estimate the number of people fed. Consistent with previous global estimates around year 2000 [3, 4], our updated analysis for year 2019 shows that synthetic nitrogen fertilizers are used to produce half of the global protein supply, feeding 3.8 billion people worldwide (figure 1).

Zoom In Zoom Out Reset image size

According to the FAO, China is the largest consumer of synthetic nitrogen fertilizers (27 Mt nitrogen, Mt N, per year), followed by India (19 Mt N per year), the United States (12 Mt N per year), Brazil (5 Mt N per year), Pakistan (4 Mt N per year), Indonesia (3 Mt N per year), and Canada (3 Mt N per year) [26]. Yet, by considering country-specific nitrogen use efficiencies and nitrogen intakes in diets, we find that India is the country producing proteins that feed the largest number of people with synthetic nitrogen fertilizers, 646 million people, followed by China (530 million people), the United States (480 million people), and Indonesia (225 million people) (figure 1). These results quantify the number of people fed by synthetic nitrogen fertilizers and the reliance of global food production on synthetic nitrogen fertilizers. Because of higher inefficiencies in fertilizers applications and nitrogen intakes in diets, China—the largest consumer of synthetic nitrogen fertilizers—can feed less people than India—the second largest consumer of synthetic nitrogen fertilizers. We also show that the Unites States produces enough protein to feed 480 million people, a number larger than its population. This is because the Unites States are a major exporter of food and a fraction of proteins produced from synthetic nitrogen fertilizers within the country are internationally traded.

International trade is increasingly transporting commodities from one country to another and exacerbating vulnerability to supply shocks [53]. About one fourth of global food is traded among countries [54]. Similarly, synthetic nitrogen fertilizers are produced and traded around the world [6]. In 2019, synthetic nitrogen fertilizer exports equated 38% (47 Mt N per year) of global production [26]. We analyze country-specific trade balances to quantify synthetic nitrogen fertilizers self-sufficiency and determine net-importers and net-exporters. Using country-specific data of fertilizers production and use [26], we calculate the net-imports and net-exports of nitrogen fertilizers.

Our analysis shows that exports of nitrogen fertilizers are concentrated in a few countries (figure 2(A)). Russia is the largest net-exporter of synthetic nitrogen fertilizers (9.2 Mt N per year), followed by China (5.6 Mt N per year), Egypt (3 Mt N per year), Qatar (3 Mt N per year), Saudi Arabia (2.6 Mt N per year), the United States (1.6 Mt N per year), Oman (1.6 Mt N per year), and Canada (1.4 Mt N per year) (figure 2(A)).

Zoom In Zoom Out Reset image size

However, some net-exporters of synthetic nitrogen fertilizers, such as China, the Netherlands, Germany, Belgium, South Korea, Poland, and Japan are large importers of natural gas [55], which make these countries vulnerable to energy shocks—as highlighted by the current energy crisis [56, 57]. To investigate the interplay of food and energy supply, we then investigate the self-sufficiency of all countries worldwide when accounting for the share of synthetic fertilizers produced from imported natural gas. By accounting for natural gas imports [55], we find that the number of countries that can produce synthetic nitrogen fertilizers self-sufficiently decreases (figure 2(B)). China, Germany, the Netherlands, Finland, South Korea, Japan, Belarus, and Poland become net-importers of synthetic nitrogen fertilizers (via the import of natural gas) (figure 2(B)). In contrast, major fossil fuels producers, such as Russia, Qatar, Saudi Arabia, Egypt, Oman, and Iran remain net-exporters of synthetic nitrogen fertilizers (figure 2(B)).

This analysis shows which country can produce enough fertilizers to meet domestic demand. It allows to understand the countries' vulnerability to food, fertilizers, and energy supply shocks, and it highlights the strong connection between the food and energy systems, with a handful of countries controlling either the food or the energy resources required to produce fertilizers.

The reliance of synthetic nitrogen fertilizer on international trade represents a threat to global food security and resilient food systems [53]. Food security is achieved 'when all people, at all times, have physical, economic and social access to sufficient, safe, and nutritious food to meet their dietary needs and food preferences for an active healthy life' [58]. In this study, we consider the nutrient security component of food security considering a safe physical access to proteins [59]. We quantify the number of people fed with traded synthetic nitrogen fertilizers when accounting for fertilizers trade with and without embedded natural gas imports (i.e. analysis presented in figure 2(B)).

We find that 1.07 billion people per year are fed from food reliant on fertilizers imports (figure 3). However, when accounting for embedded natural gas imports, or the share of synthetic fertilizers produced from imported natural gas, we find that an additional 710 million people rely on natural gas imports used to produce synthetic nitrogen fertilizers. Globally, 1.78 billion people per year are fed from food reliant on imports of either fertilizers or natural gas (figure 3). India is the country relying the most on nitrogen fertilizers import, with 176 million people relying on direct import of fertilizers, and 379 million people relying on either imported fertilizers or imported natural gas (figure 3). China—a country self-sufficient in terms of nitrogen fertilizers—becomes a net-importer when accounting for natural gas imports, with 151 million people being fed from fertilizers produced from imported natural gas (figure 3). When accounting for imports of either fertilizers or natural gas, nitrogen fertilizers are used to produce food that can feed 185 million people in Brazil, 93 million people in France, 64 million people in Ukraine, and 58 million people in Turkey (figure 3). Without trading of fertilizers or natural gas, food shortages will spread with devastating impacts on millions of people.

Zoom In Zoom Out Reset image size

Ammonia (NH₃) synthesis for synthetic nitrogen fertilizers production is energy and carbon intensive. The global production of NH₃ is about 183 Mt per year; ∼70% of which goes into synthetic nitrogen fertilizers, whereas the remaining fraction is used for plastics, explosive, and textile production [6]. While global greenhouse gas emissions from NH₃ production are estimated to be 450 Mt CO₂ per year [6], country-specific carbon emissions from synthetic nitrogen fertilizers production have been overlooked until recently [17]. Using country-specific data of electricity carbon footprint and energy mix for NH₃ production, accounting for natural gas leaks along the supply chain, and modeling several processes for net-zero NH₃ production, we quantify country-specific CO₂ emissions resulting from the business-as-usual production of synthetic nitrogen fertilizers.

Our analysis shows that 310 Mt CO₂ per year are emitted globally to produce synthetic nitrogen fertilizers (figure 4). China is responsible for the largest CO₂ emissions related to the production of synthetic nitrogen fertilizers (117 Mt CO₂ yr⁻¹), followed by India (45 Mt CO2 yr⁻¹), the United States (31 Mt CO₂ yr⁻¹), Brazil (11 Mt CO₂ yr⁻¹), Pakistan (8 Mt CO₂ yr⁻¹), and Indonesia (8 Mt CO₂ yr⁻¹) (figure 4).

Zoom In Zoom Out Reset image size

NH₃ production accounts for ∼90% of the total energy consumption and CO₂ emissions of the nitrogen fertilizer industry [6]. Therefore, achieving net-zero CO₂ emissions in NH₃ production would represent a major step towards net-zero fertilizers. The conventional Haber–Bosch process is a highly integrated process that can be divided into two main steps: (a) hydrogen production from fossil fuels, and (b) NH₃ synthesis from the Haber–Bosch reaction. Hydrogen and nitrogen production are energy-intensive processes that can have different carbon intensities depending on the energy source and feedstock [6]. In the business-as-usual route, fossil fuels are used as feedstock to produce hydrogen and provide the energy required for the synthesis of NH₃, hence resulting in CO₂ emissions (supplementary figure 1(A)). In most countries, hydrogen is currently manufactured through SMR of natural gas; in China, hydrogen is produced via coal gasification [6]. Although improvements in energy efficiency and carbon intensity are underway and have reduced the emission intensity of NH₃ production by 12% over the last fifteen years [6], net-zero NH₃ production needs to phase-out from fossil fuels.

Net-zero CO₂ emissions in NH₃ production can be achieved in multiple ways: through the production of hydrogen from fossil fuels integrated with carbon capture and storage [6]; through the production of hydrogen from electrolysis using low-carbon electricity [60]; and through the production of hydrogen from biomass gasification, such as wood chips from crop and forestry residues [36]. In the carbon capture and storage route (supplementary figure 1(B)), NH₃ is still produced from fossil-fuels via the conventional Haber–Bosch process [6]. Carbon dioxide emission generated during NH₃ synthesis are captured, transported, and permanently stored in suitable underground geological structures [34]. In the electrification route, hydrogen is produced from water electrolysis via low-carbon electricity, which also power the Haber–Bosch process [60] (supplementary figure 1(C)). In the biomass route, CO₂ is captured from air via photosynthesis during biomass growth and then emitted upon synthesis and disposal of the biomass-based product, thus resulting in net-zero CO₂ emissions. Biomass contains both the carbon and hydrogen atoms, as well as the energy required for the synthesis of NH₃ [44] (supplementary figure 1(D)). For all routes to reach net-zero emissions, residual emissions along the production chain must be offset with negative emissions, for example via direct air carbon capture coupled with CO₂ and storage [6].

While all net-zero routes described above are technically feasible [6], and some allow to avoid reliance on fossil fuels, a holistic approach is needed to quantify their environmental feasibility and avert unintended environmental consequences. We quantify the land, water, and energy requirements of all net-zero routes and compare them with a business-as-usual route to produce NH₃ (See Methods section for a description of all net-zero routes). All net-zero routes are more land, energy, and water intensive than the business-as-usual route (figure 5); this is the cost of achieving net-zero emissions. Overall, the biomass route is the most water- and land-intensive route (mostly due to the high water and land intensity for growing the biomass feedstock), while the electrification route is the most energy-intensive (mostly due to the amount of electricity required to produce hydrogen via water electrolysis) (figure 5).

Zoom In Zoom Out Reset image size

Table 1 reports the reference values of global CO₂ emissions, energy requirements, land use and water consumption, reiterating that achieving net-zero emissions is possible, but requires high energy, land, and water resources. For example, decarbonizing NH₃ production with the electrification route will require TWh of electricity (or 5% of year 2019 global total electricity consumption), compared to the 48 TWh currently used under a business-as-usual production pathway (table 1). Achieving net-zero emissions via the biomass route requires 26 million hectares of land and 255 km³ of water, about a factor 1000 higher than the resources required by the current production (table 1). Overall, achieving net-zero emissions in NH₃ production is technically possible. However, water, land, and energy trade-offs should be thoroughly considered to ensure a sustainable transition to net-zero food systems.

Feeding the growing and increasingly affluent human population requires doubling global food production by 2050 [61]. To meet this demand, nitrogen fertilizers are required to increase agricultural productivity. Currently, half of the world population is supported by synthetic nitrogen fertilizers, produced via the Haber–Bosch process—a carbon- and energy-intensive process that commit global food systems to rely on fossil fuels. This study sheds light on the global food security implications of current fossil fuel-based synthetic nitrogen fertilizers production, which is exposed to supply and energy shocks. It identifies producers and consumers of synthetic fertilizers and the associated number of people fed, when considering the interplay with energy supply. Furthermore, it presents routes to achieve net-zero emissions in the production of fertilizers, and investigates their potential for improving food security, as well as their impact on land and water scarcity. These results can inform the development and deployment of resilient and sustainable strategies to achieve net-zero food systems.

Our results quantify the reliance of the global food system on synthetic nitrogen fertilizers and natural gas trade. We find that 1.07 billion people per year are fed from food reliant on fertilizers imports (figure 3). However, when accounting for embedded natural gas imports, or the share of synthetic fertilizers produced from imported natural gas, we find that 1.78 billion people per year are fed from food reliant on imports of either fertilizers or natural gas (figure 3). Overall, a handful of countries control the trade of synthetic nitrogen fertilizers, either via their production capacity or energy (mostly natural gas) resources. These countries are Russia, Egypt, Qatar, and Saudi Arabia. Multiple countries, namely China, Germany, the Netherlands, Finland, South Korea, Japan, Belarus, and Poland, are net-importers of synthetic nitrogen fertilizers via natural gas import, despite their fertilizer production capacity (which could make them net-exporters of nitrogen fertilizers if they had more fossil fuels resources) (figure 2). This unveils complex risks and threats to global food security, with more than 20% of the global population being exposed to either supply shocks via fertilizers trading or energy shocks via embedded energy feedstock (mostly natural gas) trading.

Our results show that imports of synthetic nitrogen fertilizers and embedded natural gas produce food that can feed 379 million people in India, 185 million people in Brazil, and 151 million people in China (figure 3). While Brazil and the United States are major food exporters and nitrogen fertilizers are used to produce food that feed people in other countries, in India and China most of the food produced from fertilizers is used to feed domestic population, making domestic food production particularly reliant on trade and energy and supply shocks.

Being more energy-intensive (figure 5), net-zero NH₃ production will not necessarily decrease vulnerability to energy shocks. For example, net-zero NH₃ production based on the electrification route could reduce the vulnerability to shocks on commodity markets, at least for what pertains oil, methane, and coal, but would be still vulnerable to electricity prices.

Transitioning fertilizer production to net-zero CO₂ emissions can have the twin benefit of reducing CO₂ emissions while enhancing food security. Indeed, the deployment of processes to synthesize nitrogen fertilizers from renewable energy (i.e. electrification and biomass route) can reduce CO₂ emissions while averting reliance on imports of fossil fuels. In contrast, whereas carbon capture and storage allows to achieve net-zero emissions, it does not reduce the reliance of the food system on fossil fuels; it still uses an average of 77 Mt of carbon from fossil fuels per year, hence making the global food system vulnerable to energy shocks. In addition, carbon capture and storage would require a widely spread infrastructure to transport [35, 62] and permanently store the CO₂ captured at the production site [63]. Whereas recent assessments indicate that a vast storage capacity might be available, i.e. between 7000 and 55 000 Gt CO₂ can be stored worldwide [64], CO₂ storage still faces issues concerning the actual availability, accessibility, and acceptance of storage sites [65, 66]. An alternative option to geological CO₂ storage is carbon dioxide mineralization, which consists in reacting CO₂ with metal cations to form stable carbonate materials and achieve permanent CO₂ sequestration [67–69]. Importantly, our assessment accounts for natural gas leaks along the supply chain, here assumed to be 1.5% of the required natural gas [50]. However, it is worth noting that natural gas leaks affect carbon emissions, hence land, energy, and water consumption required to achieve net-zero emissions of the carbon capture and storage route only. Arguably, the carbon capture and storage route would find a better use for carbon-rich chemical products, such as methanol and plastics, which, contrary to NH₃, contain the carbon molecule in the final product [22].

Contrary to carbon capture and storage, the electrification and biomass routes can achieve net-zero emissions while avoiding using fossil fuels. However, the electrification route would require twenty-five times more energy than the business-as-usual route. The biomass route would require one thousand times more land and water than the business-as-usual route, using 26 million hectares of land and 255 billion cubic meter of water (table 1). To grow this vast increase in biomass, further nitrogen inputs [70], and transport and processing facilities would be required. In addition, both biomass and electricity will be required to achieve net-zero emissions in other sectors and a competition for these limited resources could constrain their use for NH₃ production [71]. To avert unintended environmental consequences on natural resources and biodiversity and additional land, water, and fertilizers use, biomass should be sourced sustainably from waste biomass, forestry residues, and crop residues [72, 73].

While net-zero nitrogen production routes could solve energy and food security issues present in business-as-usual NH₃ production, these alternative NH₃ production routes could create inequalities in NH₃ nitrogen fertilizers production with more technically advanced economies continuing to dominate production.

By emitting 310 Mt CO₂ per year (figure 3), business-as-usual NH₃ synthesis for synthetic nitrogen fertilizers production commit humanity to emissions levels not compatible with the net-zero targets required to keep global warming below 1.5 °C [74]. An additional 30 Mt CO₂ per year are estimated to come from NH₃ transport [17]. Although NH₃ is not a greenhouse gas, its overuse lead microbes in the soil to convert it into nitrous oxide, a greenhouse gas three hundred times more powerful than carbon dioxide and responsible for stratospheric ozone depletion [75, 76]. It is estimated that nitrogen fertilizers emit 2.3 Mt of nitrous oxide per year, equivalent to 670 Mt CO₂ emissions per year [6], bringing global total emissions (direct and indirect emissions) from synthetic nitrogen fertilizers to 1010 Mt CO₂ per year when accounting for emissions from NH₃ synthesis for nitrogen fertilizers, or 2% of global greenhouse gas emissions.

Economic and population growth are expected to double global food demand by 2050 [61]. Thus, synthetic nitrogen fertilizers are envisaged to continue to be a major and growing component of agricultural productivity in the 21st century [5, 77, 78]. While the net-zero routes analyzed here can abate emissions on the supply-side, demand-side measures can reduce future NH₃ demand and significantly ease the task of achieving net-zero emissions while considering environmental trade-offs and socio-political shocks (e.g. related to food and energy supply) [79, 80]. Encouraging diets with low nitrogen footprint or less meat, reducing food losses and waste, and improving nitrogen use efficiencies can reduce future NH₃ demand [15]. First, global average nitrogen use efficiency—the share of applied nitrogen incorporated in food production—is estimated to be around 46%, meaning that more than half of synthetic nitrogen is dispersed in the environment and not used to grow crops [5]. Precisions agriculture can increase the efficiency of nitrogen fertilizers application to crops [81]. Second, nitrogen losses from farm to fork are estimated to be between 41% and 44% and are mainly due to harvesting and distribution losses, and food waste [3, 29]. Such losses can be reduced by reducing food waste and improving efficiencies in food supply chains. Third, a dietary transformation to less nitrogen-intensive diets can reduce nitrogen demand. While the recommended daily protein intake for a healthy diet is estimated to be ∼50 g per person per day (or about 9 g of nitrogen per person per day), today the global median intake is 84 g of protein per person per day [26]. Moderating the consumption of animal-based food can reduce nitrogen demand [82, 83]. Importantly, while average protein intake is above the recommended value for healthy diets, one billion people still suffer from protein deficiency worldwide [84], indicating inequalities in our food systems.

Considering losses, inefficiencies, and waste [85], it is estimated that only ∼20% of produced synthetic nitrogen fertilizers feed global population [3, 5, 29, 33]. Therefore, ∼80% of synthetic nitrogen fertilizers produced via the Haber–Bosch process are lost due to inefficiencies in our food systems. We believe that influencing consumer behavior to eat less animal products, improving nitrogen fertilizers use efficiencies, and reducing food waste and losses, is key to achieve net-zero emissions.

Transitioning from a linear to a circular economy that capture and recycle nitrogen from waste can moderate use of resources and energy required to produce synthetic nitrogen fertilizers [6]. Promoting the use of organic fertilizers such as manure or compost can reduce NH₃ demand [86]. Animal manure nitrogen outputs are a major source of nitrogen recovery and recycling globally [6]. The digestate produced from anaerobic digestion of livestock manure can be spread over croplands to recover nitrogen [32]. However, organic fertilizers are often more expensive, slower in releasing nutrients, and are not presently capable of supporting the demands of current or future generations [6, 9, 87]. There are also promising scientific developments underway for alternative fertilizers, but many of these approaches need further development. Bioinformatics and plant genomics can both reduce fertilizer usage [88]. Electrochemical synthesis and plasma activated processes are other promising approaches that could be deployed as an alternative to the Haber–Bosch process to produce NH₃ [89–95].