Can our global food system meet food demand within planetary boundaries?

https://doi.org/10.1016/j.agee.2017.06.001Get rights and content

Highlights

  • A tool is built to explore the impact of food demand on the planetary environment.

  • The entire food system is analysed by linking multiple measures to multiple goals.

  • Improvements are necessary to protect the planetary environment towards 2050.

  • Combining measures is essential to stay within planetary boundaries.

  • Better underpinning of the boundaries for N and P losses is needed.

Abstract

Global food demand is expected to increase, affecting required land, nitrogen (N) and phosphorus (P) inputs along with unintended emissions of greenhouse gasses (GHG) and losses of N and P. To quantify these input requirements and associated emissions/losses as a function of food demand, we built a comprehensive model of the food system and investigated the effects of multiple interventions in the food system on multiple environmental goals. Model outcomes are compared to planetary boundaries for land system change, climate change and the global N and P cycles to identify interventions that direct us towards a safe operating space for humanity. Results show a transgression of most boundaries already for 2010 and a drastic deterioration in the reference scenario for 2050 in which no improvements relative to 2010 were implemented. We defined the following improvements for 2050: reduction of waste, less consumption of animal products, higher feed conversion efficiency, higher crop and grassland yields, reduction of N and P losses from agricultural land and reduction of ammonia (NH3) volatilization. The effects of these measures were quantified individually and in combination. Significant trade-offs and synergies in our results underline the importance of a comprehensive analysis with respect to the entire food system, including multiple measures and environmental goals. The combination of all measures was able to partly prevent transgression of the boundaries for: agricultural area requirement, GHG emission and P flow into the ocean. However, global mineral N and P fertilizer inputs and total N loss to air and water still exceeded their boundaries in our study. The planetary boundary concept is discussed in relation to the selected variables and boundary values, including the additional necessity of eliminating the dependency of our food production on finite P reserves. We argue that total N loss is a better indicator of the environmental impacts of the global N cycle than fertilizer N input. Most measures studied in this paper are also on the agenda of the United Nations for Sustainable Development, which gives added support to their implementation.

Introduction

Food production relies on the availability of resources, such as land, fresh water, fossil energy and nutrients. Current consumption or degradation of these resources exceed their global regeneration rate (e.g. Molden, 2007, Bindraban et al., 2012, Van Vuuren et al., 2010). Furthermore, food production is associated with emissions that deteriorate earth’s environmental quality (e.g. Rockstrom et al., 2009), like nutrient leaching that causes eutrophication of natural ecosystems, and greenhouse gas (GHG) emissions that contribute to global warming (IPCC, 2014).

To guide human activities, Rockstrom et al. (2009) and Steffen et al. (2015) selected control variables for nine critical earth system processes and assigned environmentally safe planetary boundaries to these variables. These boundaries refer, inter alia, to the following control variables which are closely related to the use of resources in agriculture and subsequent effects of their use: (i) the area of forested land, (ii) the energy imbalance (or total radiative forcing), (iii) industrial and intended biological nitrogen (N) fixation, (iv) the flow of phosphorus (P) from fertilizers to erodible soils and (v) the flow of P from freshwater systems into the ocean (Table 1). Forested land (i) was selected because of its assumed stronger role in land-climate interactions compared to other biomes. For example, evapotranspiration from the land surface may change when tropical forests are cleared, and boreal forests influence the albedo of the land surface. The energy imbalance (ii) integrates factors that affect the earth’s energy balance such as the atmospheric concentrations of greenhouse gasses, and is closely linked to global warming. Industrial N fixation refers mainly to the production of N fertilizers for agricultural use, whereas intended biological N fixation is linked to the cultivation of leguminous crops (iii). Together, they have dominated the recent increase of reactive N input on earth during the last decades. P fertilizer flow to erodible soils (iv) causes eutrophication of watersheds according to Steffen et al. (2015). They assumed that all cropland soils are in principle erodible, and that P addition to watersheds is primarily linked to P fertilizer application. P flow from rivers into the ocean (v) was selected to prevent large-scale ocean anoxic events with detrimental effects on biodiversity.

Steffen et al. (2015) provides detailed information on above variables including their motivation for the planetary boundaries as given in Table 1. According to Steffen et al. (2015), the current states of above five control variables are already transgressing their safe zones, i.e. pushing the earth from its relatively stable state of the Holocene into the Anthropocene with unforeseen consequences. The expected growth in world population and increase of the animal protein share in the human diet will probably cause further pressure on agricultural resources and the environment, and may lead to further exceedance of related boundaries.

Several research groups modelled (parts of) the global food system in relation to resource use and emissions to gain insight into feasible options of producing sufficient food within the safe operating space of our planet. For example: land requirement by Erb et al. (2016); GHG emissions by Wollenberg et al. (2016); cropland N cycle by Zhang et al. (2015), Billen et al. (2015) and Liu et al. (2010); whole N cycle by Bodirsky et al. (2014); cropland P cycle by MacDonald et al. (2011) and Sattari et al. (2016); whole P cycle by Van Vuuren et al. (2010); whole N and P cycles by Bouwman et al. (2013) and Sutton et al. (2013). These publications describe only one or two resources or a part of the entire food system, and their results are not always based on a quantitative model where inputs and outputs are linked. However, we need a quantitative description of the main inputs and outputs of the food system and their interdependencies to understand the effects of our food system on multiple earth system processes as in Table 1. We hypothesize that such an integral system description will offer new insights because of unforeseen trade-offs and synergistic processes, especially if multiple measures are applied to achieve multiple environmental goals.

The objectives of our study are 1) to quantify agricultural land use, N and P fertilizer requirements, GHG emissions and losses of N and P for a number of food system scenarios within the context of planetary boundaries and 2) to contribute to the global debate on whether and how our food system can function within planetary boundaries. To our knowledge, such a quantitative analysis has not yet been published, linking current and future food demand to required input levels of land, N and P fertilizers and to associated emissions of GHG, N and P. Therefore, we developed a comprehensive model by integrating all major food and resource flows from the actual food intake by the human population, via food processing, livestock production, up to the N and P balance of cropland and permanent grassland soils. Feedback processes are included in the model to obtain a more realistic calculation.

This paper presents the methodology and the results of a study that explores the consequences of food demand as well as the mitigating effects of potential improvements in agricultural production methods and food chain adjustments. Our analysis is as yet limited to the global food system of 2010 and a projected food demand for 2050. However, the methodology can easily be updated with other data (both spatial and temporal) to explore, for instance, differences among regions or recent developments in food systems.

Section snippets

Overview

We developed the model BIOSPACS (Balancing Inputs and Outputs for the Sustainable Production of Agricultural CommoditieS) to quantify N and P flows between five interacting components in the food system and those across the system’s boundary as a function of food demand (Fig. 1). Related agricultural land requirements and GHG emissions from agricultural production are also calculated. The five components are: (1) human Population consuming food items, (2) the Food balance supplying (non-)food

Required land area and yields

Total required area calculated by BIOSPACS is slightly less than total available agricultural land in 2010 (Table 3), which agrees with data from FAOSTAT. Because yields in the reference scenario for 2050 are equal to the yields of 2010, required grassland and cropland areas increase by 78% and 67%, respectively. This is more than the increase in total food supply, viz. 60%, due to the higher share of animal products in the human diet of the reference scenario for 2050.

Roughly the same amounts

Future food demand

In our study we use the projections of food demand and diet change by 2050 from Alexandratos and Bruinsma (2012). These were reviewed together with results from 10 other demand modelling approaches by Valin et al. (2014) who compared food demand changes between 2005 and 2050 for their reference scenario. Large differences were found among the projections of these global economic models, viz. increases of 59–98% for total food and 61–144% for food from livestock. These ranges illustrate the

Concluding remarks

In 2015 world leaders adopted a number of Sustainable Development Goals (SDG, https://sustainabledevelopment.un.org/sdgs) and several targets for each goal. These targets are linked to our study, such as target 12.2: to achieve sustainable management of natural resources, 12.3: to halve global food waste per capita, 14.1: to reduce marine pollution, in particular from land-based activities and 15.2: to halt deforestation. Tools like BIOSPACS can be used to quantify how specific targets per

Acknowledgements

This research was funded by the International Fertilizer Development Center and the Dutch Ministry of Economic Affairs [KB-30-004-007]. The authors thank Ben Rutgers for his assistance with processing data from FAOSTAT.

References (30)

  • K.H. Erb et al.

    Exploring the biophysical option space for feeding the world without deforestation

    Nat. Commun.

    (2016)
  • FAO

    Food Balance Sheets A Handbook

    (2001)
  • J. Gustavsson et al.

    Global Food Losses and Food Waste: Extent, Causes and Prevention

    (2011)
  • D.F. Herridge et al.

    Global inputs of biological nitrogen fixation in agricultural systems

    Plant Soil

    (2008)
  • H. Jönsson et al.

    Guidelines on Use of Urine and Faeces in Crop Production

    (2004)
  • Cited by (169)

    View all citing articles on Scopus
    View full text