Agriculture in a Hot World

Abstract

Against the background of a brief history of global agricultural production of the last 12.000 years with a focus on the last 100 years, illustrates the challenges today’s food system is confronted with given an unfortunate but likely scenario of a +3° temperature increase. Our review of the available evidence paints a quite dreadful picture. A moderate gain in the area suitable for agriculture is confronted with substantial yield losses through strains on crop physiology, multitrophic interactions, and more frequent extreme events. Self-amplifying feedbacks are unresolved and might lead to further unexpected losses. Consequently, all well-known pathways to eradicating global hunger while bending the curve of biodiversity loss need to be implemented today: Transitioning to less energy-rich diets, closing yield gaps through agroecological principles, adopting modern breeding technologies to foster stress resilience and yields, minimizing harvest losses and food waste and a trade system that serves to distribute food fairly and to compensate for possible climate-related yield losses, while locally adapted cultivation methods should be developed further and not be replaced. Efficiency gains in agriculture only, albeit indispensable, will not be enough to achieve food security under severe climate change.

As this text takes shape, there are about 8 billion people living on our planet. There is no need to go to bed hungry these days. Healthy, sufficient, and diverse food could in principle be available in sufficient quantities to everybody. In fact, global food production averages an energy equivalent of 5000 kilocalories (kcal) per person per day. Per capita consumption differs by 430% between the richest countries (Australia, Germany, Canada, …) with more than 8000 kcal per person and day and the poorest countries (Chad, Congo, Niger, …) with about 2000 kcal (Tilman et al., 2011).

At the same time, about 800 million people are still undernourished. 250 million children under the age of five are either malnourished, have reduced height growth or are significantly over-nourished (Fig. 1). Children in countries of the global South are especially affected by malnutrition. In these regions, people must get by on less than $8 per day, even though the situation has improved significantly in recent decades: Child mortality halved between 1990 and 2017, and the number of people living in extreme poverty fell to 736 million in 2015.

Fig. 1

Malnutrition has three manifestations: reduced height (blue), underweight (green) and overweight (orange). Sorted by income groups—from extreme poverty with less than $2 per person per day to the rich states with more than $32—the absolute numbers of affected children under 5 years of age and their percentage share in this population group are shown here. (Data from UNICEF, WHO)

However, there are regional deviations or setbacks: Between 2013 and 2015, child mortality in sub-Saharan Africa rose slightly again (Rosling et al., 2018). Africa is repeatedly shaken by individual regional hunger crises, such as in Madagascar in 2020. These are only two examples that show that a development that is quite positive on a global average can turn into the opposite on a regional level. It is to be feared that such crises and setbacks will become more frequent due to climate change.

Food security is not exclusively a question of production quantity. Besides stable production at a sufficiently high level, it is a question of availability, access, distribution, and use of food. This makes it clear that agricultural production and food security are very complex issues, of which only the most important aspects can be highlighted here in relation to the supply of food to the world’s population. What our food supply will look like if climate change continues unchecked is therefore not easy to answer, but nonetheless a clear (rather bleak) picture emerges.

According to UN/WHO figures, there will be between 8.8 and 11.6 billion people on the planet by 2060. These people are expected to have a higher income and their eating habits will have shifted towards more energy-rich diets. Accordingly, the demand for food will increase by a further 59–98% by 2050 (Valin et al., 2014). At the same time, climate change will cause annual average temperatures to rise and extreme events to increase in frequency and intensity (Hansen & Cramer, 2015). Climate change will lead to changes in the distribution of water, earlier flowering dates, longer growing seasons, more frequent crop failures, increased pest infestations, and much more. Yield patterns will change, and it is very likely that agricultural yields will not increase, but rather decrease, as 52% of agricultural land is already considered degraded (WWF, 2020).

To get a sense of how our food supply will evolve in the near future, we will here examine three issues: First, we have a look at the Anthropocene and show how our supply of agricultural products has improved so much that no one should go hungry today, which unfortunately is not the case. We will then, second, also look at where we come from, at what has shaped our basic beliefs about our food system: namely a process of productivity optimization of agricultural crops that has been going on since about 12,000 BC, evolving in the stable climate of the Holocene and optimized through domestication and breeding (Doebley et al., 2006). Finally, we are trying to understand the complex ecological interactions that are crucial for sufficient and stable production of healthy food. Recognizing that, if average temperatures rise by (much) more than 2 degrees, our world will look very different. Business as usual is by no means an option. Investing in innovation and optimization of food production might, on the contrary, be just one part of the story. In the end, we are probably left with the conclusion that a different, healthier way of life has many positive side effects, while at the same time, new technologies and breeding successes must be part of the solution (Löwenstein, 2017; Willett et al., 2019).

Our Recent Past in the Anthropocene

Since 1960, global food production has risen so much that, despite a growing world population, the amount of food produced per capita has increased (Fuglie et al., 2019; Fuglie, 2021). This increase resulted only partly from the expansion of the area under agricultural cultivation. This is not to ignore that increasing deforestation of tropical rainforests is primarily driven by demand for agricultural products. But the expansion of agriculture does not explain a 2.5-fold increase in production. The main reason why agricultural production has increased so much is the higher intensity of land use. The yield potential of our crops has increased greatly in the course of domestication. For example, the Green Revolution in the 1960s let to varieties that, due to their compact growth, could convert higher amounts of nitrogen fertilizer into produce (Bailey-Serres et al., 2019). On agricultural land, more fertilizer, more pesticides, but also more water are used to achieve this increased yield potential. What plants lack in the field, the so-called limiting factors, are supplied by humans.

Mitscherlich and Liebig had already developed a general law in the nineteenth century (Mitscherlich, 1909). Today we remember it with the term “Liebig’s Minimum Principle” (Liebscher, 1895). It says that yields can be increased to a maximum by eliminating all limiting factors—but not indefinitely. Mitscherlich also formulates the principle of diminishing returns: the higher the yield, the more likely it is that further increases in yield, for example by adding fertilizers, will have only a minimal positive effect on yield. So-called external effects for the environment (e.g., pollution of groundwater and the species community in the receiving water) are then all the greater. One way of reducing external effects is to increase the nutrient efficiency of crops through breeding, i.e., more efficient uptake and utilization of the given amount of nutrients. Such strategies benefit from an improved understanding of how the crop functions, such as the nitrogen uptake of rice (Hu et al., 2015). Mechanisms of wild plants, which are often highly efficient in taking up nutrients such as phosphate (Lambers et al., 2015), can also be transferred to crops (Fig. 2).

Fig. 2

Global agricultural production since 1960 in index values (all data for 1970 are normalized to 1). While total agricultural production increases by a factor of 2.8, the population grows by a factor of two. As an example of a factor that made this increase in production possible, the amount of fertilizer applied worldwide is shown, which has increased by more than 3.5 times. Over the same period, the ratio of goods produced per amount of fertilizer used (“efficiency”) has fallen continuously. The index for the integrity of ecosystems (“Living Planet Index”) shows that this is continuously decreasing. (Data from WHO, FAOStat, WWF)

Interim conclusion: the surface area on our planet is obviously limited. Humans have already anthropogenically shaped 70% of the earth’s surface, be it through the creation of pastureland, crop farming, or the construction of roads and settlements (Watson et al., 2016; Díaz et al., 2019b). The question of the extent to which yields can be further increased through further technological developments and ultimately intensification is highly topical and open: what is certain is that the answer to this question varies greatly from region to region. In regions with a severe undersupply of nutrients, even small amounts of fertilizer lead to large increases in yields, while other regions suffer from severe oversupply, for example with phosphate (MacDonald et al., 2011). A look at current discussions of agricultural production in Europe or in Germany shows that it is difficult to achieve further increases in yield and thus higher incomes for agriculture with the currently cultivated crops on the given areas, precisely because the areas and the farms are already farmed very intensively. In Europe we are already achieving maximum yields—but is this also true globally?

An analysis of time series of the production of different renewable resources answers the question of whether these time series are subject to continuous growth or whether they tend toward saturation. In the case of fossil resources, we are (still) experiencing a continuous increase in production: mankind’s hunger for energy is so great that more and more fossil fuels are being extracted, with the well-known consequence that these continue to pollute the atmosphere and lead to a further increase in CO₂ concentrations worldwide. In contrast, for many renewable resources, it is clear that the point of maximum yield increase has long been exceeded globally (Fig. 3).

Fig. 3

Results of the determination of years of maximum yield or production growth (peak year) of products from animal husbandry (orange), arable farming (green) and the use of means of production and socio-economic variables (blue). The dot shows the respective year of maximum production growth, the size of the dot corresponds to the growth rate in the respective year, see also www.ufz.de/global-agriculture. (Seppelt et al., 2014)

However, this analysis of globalized agriculture also shows the process of intensification described above. If we look at the time series of the means of production used (“resources” in Fig. 3), we see that first the rate of increase of agricultural land (the expansion) reached its peak and slows down since 1950. Then the increase in irrigated land reached its maximum in 1978, and finally, the increase in nitrogen fertilizer use peaked in 1983. Liebig’s principle can be observed well here. First, the most fertile land was put under the plough; then land use was extended to less fertile areas; and finally, the limiting factors (first water, then nutrients) were targeted, which made further increases in yield possible. In the end, the message remains: the maximum increase in yields of all renewable resources falls almost synchronously in the period from 1989 to 2008 (median: 2006). We are managing a finite planet.

However, these increases were not linear but, with the emergence of innovations like the Green Revolution, in spurts. The question arises whether there will be innovations in the near future that will bring about a new push in yield gains. Recent breakthroughs in our functional understanding of plants hold out the prospect of such a disruptive development (Bailey-Serres et al., 2019). New possibilities of genome editing will massively accelerate the further development of crops, as precise changes can be made now, for the first time, in the genome of adapted genotypes (Chen et al., 2019). It will even be possible to make wild plants usable within a few generations, which is equivalent to domestication in fast motion (Zsögön et al., 2018).

But what might be causing the slowing of production growth in many renewable resources? It is unlikely that declining consumption and demand are the causes. Although global population growth is slower, this trend is more than offset by increasing consumption due to growing prosperity and by a shift toward more energy-rich (meat-heavy) dietary patterns is taking place.

Before we come back to how sensitive the global system of agriculture is, let us look back at how we have reacted to bottlenecks in the past millennia. This should make it clear that so far, we have always found creative responses to problems: agricultural production in the Holocene actually seems to be a continuous success story. This can be encouraging. But the retrospective also makes clear how drastic the changes coming our way are and how short the time span is in which the right decisions must be made.

Agricultural Activity in the Holocene

With the Neolithic Revolution, about 10,000 years before our era, Homo Sapiens began to leave behind his hunter-gatherer stage becoming sedentary. This has brought many changes to the in these days still very small societies, through which our species has literally grown to become the most successful, or rather most influential, species on the planet. Humans consume 25% of the annually growing net primary production (Krausmann et al., 2013). No other species in Earth’s history has ever had such a high global turnover of matter, not even the dinosaurs. We are indeed in the process of (over)shaping the Earth’s ecosystem, which is why we now speak of the “Anthropocene”.

In the beginning, one of the most important innovations for us was the domestication of the first arable crops, such as emmer, barley, lentils and rice, and of the first animals (Doebley et al., 2006). This happened in China, in New Guinea, in Central America and of course in the Fertile Crescent in what is now the Middle East (see Fig. 4a). About 2600 years before our era, about 50% of all domesticated plants and animals were already known and used (Fig. 4c). After that, it took another 4000 years until this way of cultivation spread to almost the entire surface of the earth and displaced other ways of life of the genus Homo (which were still surviving then and even today). New domesticated plants such as cotton and maize arrived from Central and South America and from regions in Africa. Much of the land used today was settled and tilled by humans at that time. Until the end of the Middle Ages, agriculture continued to expand across Europe, Asia, Africa, Central and South America, new forms of cultivation such as three-field crop rotation were invented, and later four-field rotation was added as a further innovation. Industrialization made the use of machines possible, large-scale synthesis of mineral nitrogen fertilizer made agricultural work vastly easier, less physically demanding, and thereby facilitated the cultivation of ever larger areas. By now, we have modified almost all the fertile land on the planet in some way. Since 1950, productivity has been further optimized through breeding of more efficient varieties to genetically modified crops and the development and application of sophisticated crop protection products. The application of nitrogen, phosphate, and pesticides multiplied between 1960 and 1990 (Fig. 2) (Foley et al., 2011).

Fig. 4

Agriculture developed (Map a) about 10,000 years ago in the fertile areas of the Middle East, China, but also Central America and the mountainous regions of South America (agricultural cultures of the Incas). During this time, the climate was surprisingly stable, and temperatures fluctuated by only a few degrees—a pattern that, in the last 100 years or so, we seem determined to leave behind (b). Many animal and plant species were quickly domesticated and cultivated (c). The expansion of the genus Homo over the subsequent 10,000 years has led to the worldwide spread of agricultural production methods, along with the associated cultivated species, which were (and are) not always adapted to their new locations. (Data from Ellis et al., 2010, 2013; Seppelt et al., 2014; Kaufman et al., 2020)

In the relatively stable climate of the Holocene, these innovations spread wherever possible. At the same time, homogenization took place. Since the conquest of the world in the late Middle Ages, an exchange of agricultural cultures took place, with advantages and disadvantages. On the positive side, crops that were easy to cultivate became widespread. Cereal crops from the Middle East conquered the world, as did the “American” potato and maize. On the negative side, introduced alien species have drastically changed entire ecosystems. Many species that had no natural enemies until the appearance of humans were hunted to extinction (Díaz et al., 2019a, b). The main crops cultivated by humans are 1-year grasses: cereals such as wheat, rye, barley, rice, maize, millet, etc. This homogenization has led to a one-sided and uniform production system in parts of the world (Khoury et al., 2014, 2016).

Innovation is thus a guiding principle throughout our entire development. In the course of the Holocene, the model of cattle breeders and arable farmers prevailed—and could be optimized because environmental conditions, above all a stable climate, allowed it. Of course, there were always setbacks. The decline of the Fertile Crescent, for example, was a result of both overuse and climatically induced changes in water availability. Between the fifteenth and seventeenth centuries, the so-called Little Ice Age—although it reduced the average temperature by merely 0.1 °C—brought Europe repeated crop failures due to cooler summers. But the bottom line remains the same: A stable climate and human ingenuity have created an extremely successful system of global resource use, which is now slowly reaching its production limits.

Agriculture—Still a Very Sensitive System

Even though a drastic homogenization of production took place, especially with the onset of the modern era, agricultural methods and practices are exceedingly diverse around the world—especially when one takes into account not only biophysical and climatic aspects, but also socio-economic ones.

Yield Gaps

Identifying yield gaps is one way of finding out to what extent yield increases can be achieved in the future and in which regions this is most likely to be feasible. All locations where agriculture can be practised are compared in terms of their biogeographical and climatic factors and grouped into categories. Within these categories, the locations where yields are highest are identified, and it is then assumed that this yield can also be achieved at all comparable locations, which thus are presumed to have yield gaps. The values given in the literature vary, but it can be assumed that, under current climatic conditions, the closing of all yield gaps would raise aggregate yield by 58%. If the option of growing other, more efficiently producible crops in some locations is included, the potential yield increase rises to 148% (Mauser et al., 2015). But how easy (or difficult?) is it to realize these gains in practice? Do we just have to manage agricultural land more intensively, for example, by applying higher amounts of fertilizer, and everything will be fine?

It is obvious that in some regions the necessary means of production are not available or cannot be financed. This becomes clear as soon as social and economic factors are taken into account, resulting in so-called land use systems (also called “land system archetypes” or LSA) (Fig. 5).

Fig. 5

World map of different land system archetypes (LSA). As a result of an analysis of biophysical, climatic, and socio-economic data, twelve characteristic archetypes can be distinguished, which account for between 0.1% (urban areas) and up to 20% (boreal regions in the east) of the global land area. (Václavík et al., 2013)

Land Use Systems or Land System Archetypes (LSA)

To better assess the effects of land use and to understand the diversity of land use systems around the world, a world map of land use was created and more than 30 indicators on agriculture, environment, climate, and the socio-economic situation were evaluated for this purpose.

Figure 5 (and the corresponding legend) shows for various regions whether maximum intensive cultivation is already practiced (LSA 10) or agriculture can still be intensified (LSA 7)—as well as what climatic conditions are expected and whether more or fewer people will need to be supplied in the future. In large parts of China, India, and Eastern Europe, for example, “traditional agriculture with high labor input” (LSA 7) is practiced. These areas are important because they are classic yield gap regions, meaning areas where farmers could still increase yields because little fertilizer is being applied, the infrastructure is poorly developed, or agriculture still receives little state support.

These findings also support evidence-based conclusions on what specific measures can be taken to counteract the negative effects of land use. In parts of Latin America and Southeast Asia, for example, soil erosion is extremely high (LSA 2). Since agriculture plays an important role in the national economy, the ecological problem becomes a socio-economic one and should be remedied with the highest priority. This would not only remedy environmental harms but also increase yields and returns from agriculture.

• 1. Forest systems of the tropics (14%)

  • high biodiversity,

  • increase in arable and pasture land,

  • comparatively high climate anomalies

• 2. Degraded forest and pasture systems of the tropics (0.35%)

  • high erosion

  • high share of agriculture in GDP

  • low political stability

• 3. Boreal forests of the western world (14%)

  • high GDP and political stability

  • poor accessibility

  • low land use intensity

• 4. Boreal forests of the eastern world (20%)

  • like (3), but low political stability

• 5. Urban agglomerations (0.1%)

  • well above-average population density, rising total population

  • a wide range of environmental conditions

• 6. Rice cropping systems with high yield potential (1%)

  • high proportion of arable land and high species richness

  • agriculture accounts for a large share of GDP

  • high population density

• 7. Traditional agriculture with high labor input (11%)

  • large and increasing share of arable and pasture land

  • agriculture accounts for a large share of GDP

  • good accessibility

  • a wide range of environmental and climatic conditions

• 8. Grazing systems (13%)

  • above-average and increasing share of pasture land

  • agriculture has a large share in GDP

  • below-average population density, rising population figures

• 9. Irrigation field cultivation (2%)

  • far above-average share of arable land and associated energy use

  • high rice yields

  • agriculture accounts for a large share of GDP

  • well above-average population density, rising population figures

• 10. Intensive agriculture (5%)

  • well above average, but decreasing share of arable land

  • far above-average energy use (fertilizer, pesticides)

  • agriculture accounts for a rather small share of GDP

  • good accessibility and political stability

  • temperate climate

• 11. Marginal land in developed countries (9%)

  • High GDP and low population density

  • Little to no arable yields, but some pasture land

• 12. Deserts and wastelands in the countries of the Global South and emerging economies (11%)

  • high temperatures, little precipitation

  • little to no pasture and arable land

  • low GDP

Biodiversity—A Threatened Production Factor

The question how we can feed all now and in the future is often answered by saying that a higher intensity of cultivation leads increased production. Irrespective of whether this is feasible everywhere, more intensive land use (and expansion into still semi-natural ecosystems) has negative effects, even though increased application of production inputs such as fertilizers or chemical pesticides did facilitate enormous gains in global agricultural production.

A critical effect of this is the clear downward trend in biodiversity indicators: the so-called Living Planet Index has declined by about 25% since 1970. The Intergovernmental Platform on Biodiversity and Ecosystem Services (IPBES) states in its first global assessment (Díaz et al., 2019a) that our way of land use, from the intensity of resource use to the homogenization of landscapes, is the key driver of biodiversity loss. If we continue on this path, we can expect the extinction of 500,000 to one million species can be expected (Díaz et al., 2019a, b).

We cannot answer the question of possible future scenarios for agriculture under a significantly warmer climate without also considering the aspect of biodiversity. Agricultural production depends on functioning ecosystems, and these are not possible without maintaining biodiversity. Biodiversity is a crucial agricultural production factor: birds and insects fight potential pests (herbivores) and thus help stabilize yields. About 70% of all crops depend on pollination services, which in turn are provided by insects and birds or bats. Almost all agricultural production is based on sufficiently deep and fertile soils, which in turn are due to the interaction of fungi and microorganisms. Currently, soil degradation has reduced the productivity of 23% of the world’s land area, and the loss of pollinators threatens $235–$577 billion worth of annual global crop yields (Díaz et al., 2019a).

Intensification is therefore a double-edged sword: it eliminates limiting factors and thereby often leads to a yield increase (Beckmann et al., 2019), but it also decreases biodiversity (Newbold et al., 2015), thereby jeopardizing yields in the long run. This is critical in agricultural systems that have been used rather intensively for a long time (especially in archetypes LSA 6 and 7, Fig. 5).

Here, yield increases of up to 80% are contrasted with a decline in biodiversity of up to 30%, and we recognize that such systems are highly sensitive and do not react linearly (Fig. 6). Therefore, it does not seem wise to pursue agriculture economically optimised at the stress limit, i.e., to clear landscapes to the maximum and to manage them with maximum intensity—especially when one considers that climate change will increase disturbances and weather extremes. We should therefore aim to achieve stable crop production at a sufficient level with a reduced supply of energy and other inputs. Crops that are pathogen-resistant and make better use of nutrients can contribute to this. Through more diverse, small-scale land management and the continuous development of crops adapted to changing conditions, agro-ecosystems and biodiversity can be preserved and yields stabilized (Egli et al., 2020, 2021a, b).

Fig. 6

Taking into account that biodiversity is a production factor, agricultural yields can decline again when a certain intensity of land use is reached (left). Biodiversity positively influences the level and stability of yields but is negatively influenced by the intensity of land use or management (right). This is especially true in cases where crops are sensitively dependent on functioning ecosystems or a stable high level of biodiversity and technical substitutions are difficult to achieve. This trade-off has been shown in a meta-analysis for forestry, arable and pasture systems, and various intensification measures. (Graphics adapted from Beckmann et al., 2019; Seppelt et al., 2020)

What Is Efficient Agricultural Production?

Conventional intensification is also questionable in other respects. This is due to how “efficiency” is understood. The basic idea of agricultural production is (or rather was) to use the fertility of the soil to convert solar energy into biomass via photosynthesis. While this led to a (net) gain in energy for centuries, conventional agriculture is now a loss-making business in terms of energy. A measure of this relationship is so-called Total Factor Productivity (TFP) which, due to anthropogenic climate change, has deteriorated by 21% since 1961 (Ortiz-Bobea et al., 2021). For countries in Africa and Latin America (incl. The Caribbean), TFP has even declined by 26–34%. Taking increasing yields as a given, this means that more and more energy, labor and effort have had and will have to be expended on further yield increases. Agricultural production has never been as inefficient as it is today with respect to energy balances (Fig. 2).

In parallel, nutrient withdrawals have never been so high. Never before have we used so many synthetic substances which interact in correspondingly diverse and uncontrollable ways. We have moved far away from the kind of circular economy that was customary in particular through the classical combination of animal and plant production. Large quantities of plant biomass are used as feed for meat production; global meat consumption has risen apace and far exceeds World Health Organization (WHO) recommendations (Willett et al., 2019). A lower-meat diet would not only be healthier, but also more energy-efficient, free up land for food production for humans (without the diversions via animals) and reduce greenhouse gas emissions.

Small-scale agriculture continues to be the main contributor to food security (Tscharntke et al., 2012). Numerous studies show that larger farms are more productive than smaller ones, especially when looking at labor productivity. However, small-scale, sustainable agriculture that relies on polyculture can produce more food per unit of land (Deolalikar, 1981). It is obvious that small-scale farming will prevail wherever there is a lack of energy (as contained in synthetic fertilizers, for example) combined with abundant labor, as in China (Li et al., 2020) or in Andhra Pradesh’s zero budget natural farming (Grefe, 2020). In subsidy systems based on the area under cultivation, still prevalent in the European Union (Pe’er et al., 2019), such forms of farming are not even considered.

Agriculture Under a Changed Climate

What would be the effects on agricultural production if the climate protection targets of the Paris Climate Agreement were missed and the temperature rose by, for example, 3 degrees?

Due to the complexity of the issue, we must be pragmatic and simplify things. A possible temperature increase of 3 degrees is the result of different scenarios. The “worst-case” scenario RCP 8.5 of the IPCC presumably reaches this temperature increase already in 2060, whereas it would “only” occur in 2100 in the scenarios RCP 6 and RCP 4.5 (Deutsch et al., 2018). A warming by 3 degrees can therefore become real sooner or later; it will also do so at different times in different places—and there will be very many regions that will heat up by even more than the global average increase.

As we have seen in our considerations of the past, the following aspects of the socio-ecological system “agriculture” are relevant:

• the available area,

• the physiology of arable crops (growth),

• the eco-factors of an agricultural landscape, and

• the people cultivating these areas.

Sounds complicated, and it is, especially if we expand our question to ask whether future food production will suffice to feed everyone. We must then take two additional unknowns into consideration:

• the number of people—the WHO predicts 8.8–11.6 billion people for 2060 and 7.3–15.6 billion for 2100, and

• our future dietary habits.

Available Area and Possible Production Increases

Various studies have looked at potential changes in arable land under rising temperatures. Based on model calculations, it could be shown that for maize in North America and Europe, land gains of 10–20% can be expected under a warmer climate, whereas for Africa, South America and Oceania, losses of up to 40% are to be expected (Ramirez-Cabral et al., 2017). A global analysis of the basic suitability for the cultivation of agricultural crops shows a similarly uneven picture and predicts an average land gain of about 3%, whereby the suitability of newly acquired land is described as rather moderate (Zabel et al., 2014). Expansion agriculture will happen in Northern latitudes, characterized by shorter day length. Some expansion of cropland into previously unusable areas will also be possible through the development of drought-adapted or salt-tolerant crops (Fig. 7a).

The extent to which an increase in the CO₂ concentration has a positive effect on plant growth has been explored at many agricultural test stations (Fig. 7c). Especially for plants with less efficient photosynthesis (so-called C3 plants) of the mid-latitudes, such as soy, cassava, rice, potatoes and wheat, a fertilization effect and thus yield increases of 10%, 20%, and 25%, respectively, were observed (Bishop et al., 2015). But these increases can be achieved only if other potentially limiting factors, such as water and nitrogen, are available in sufficient quantities. The atmospheric CO₂ concentrations in these experiments were, however, higher than necessary for a 3 degree warmer world. In more realistic experimental set-ups, there would be significant interactions with parameters such as the availability of nitrogen and water, and clear conclusions would then be more difficult to draw. Within the framework of a model study that assumes a drastic climate scenario with a temperature increase of 6 degrees by 2100 (RCP 8.5), it can be shown for wheat, soy, and rice that a positive CO₂ fertilizer effect is offset by negative effects such as water scarcity, higher ozone levels, or the occurrence of extreme events—with great geographical heterogeneities (Lombardozzi et al., 2018) (Fig. 7e).

Growth of Agricultural Crops in a Changing Climate

Higher temperatures lead to faster growth and earlier maturity of plants. But such a shortened growing season also means that the plant has less time to accumulate biomass, and yields are reduced. The increase in extreme weather events continues to lead to more frequent crop failures. Against this background, it could be shown that the climatic changes already observed between 1980 and 2008 led to a reduction in harvest volumes for maize and wheat by 3.8% and 5.5%, respectively, most strongly in China, Brazil and Russia (Lobell et al., 2011). For wheat, rice, maize and soybean, which provide two-thirds of global calorie demand, losses of 3–7% per degree of temperature increase can be expected for the varieties available today (Zhao et al., 2017). However, 7–18% of such yield losses could be prevented with optimized sowing times alone, as a model simulation for maize, soybean and spring wheat has shown (Deryng et al., 2011).

Less studied are crops that do not belong to the main crops: vegetables, legumes, fruits. For these crops, CO₂ fertilization effects could lead to yield increases of about 22%—but also to yield losses of 9% due to increased ozone concentrations, of 35% due to water scarcity, and of 32% if the temperature were to rise by 4 degrees (Scheelbeek et al., 2020). In addition, these crops depend on intact ecosystem functions even more than annual grasses such as cereals. Pollination, which is so important, could be threatened by temperature-induced shifts in flowering times or changes in insect population dynamics (Fig. 7b).

Breeding new varieties could be another method of choice to achieve higher yields, but no yield increases beyond 1% were observed for wheat, rice, and maize between 1980 and 2008 (Fischer & Edmeades, 2010). If one extrapolating a (quite optimistic) 1% breeding-induced annual yield increase from 2020 to 2060 (at +3 degrees in the RCP 8.5 scenario), this could translate into 48% higher yields. Such a breeding success would only be realistic, however, if it also aimed at increased resistance to pests and robustness against water shortages and higher temperatures (Tester & Langridge, 2010). The fact that every non-extremophilic organism suffers from absolute temperatures of more than 42 °C is a biological constant that cannot be revised even by the most skillful breeding.

Interactions in the Agroecosystem

A warmer climate is in principle good for all species whose activity (metabolism) depends on temperature. Especially ectothermic organisms that cannot regulate their body temperature themselves (e.g., insects) will clearly benefit from a warmer climate. In our context, this is especially relevant for herbivores (plant-eating pests) (Fig. 7d).

In insects, higher temperatures lead to an increased metabolic rate: they simply eat more. Secondly, higher temperatures lead to increased offspring. Based on these basic ecological principles, there is a temperature-related increase in the activity of pests on rice, maize, and wheat, which can lead to crop losses of 10–25% per degree of temperature increase (Deutsch et al., 2018). Even in a 2-degree warmer world, wheat yield losses of about 18% for Europe and North America and of about 17% for East Asia would be predicted. Rice would suffer losses of around 59% (in South and Southeast Asia) and 32% (East Asia). For maize, the losses would be highest in North America, with up to 32%, and in Europe they would be around 23% (Deutsch et al., 2018).

Lastly, plant diseases are expected to increase in a warmer climate, by up to 25% due to weather conditions. Species introduced in the context of crop replacement could be especially affected (plus 56%) (Anderson et al., 2004). Concrete figures on possible yield losses are not yet available (Fig. 7d).

Fig. 7

Overview of possible individual effects of higher global temperatures on available area (a) and on yields due to temperature-induced changes in physiology (b), CO₂ fertilization effects (c), herbivory (d) and extreme temperatures (e) for different crops. The figures given in the text are in each case mean values of the expected changes under a warmer global climate with a 3 degrees higher mean temperature (the error bars show the existing uncertainties). (Data from Deryng et al., 2011; Zabel et al., 2014; Zhao et al., 2017; Ramirez-Cabral et al., 2017; Deutsch et al., 2018; Jägermeyr et al., 2021)

The Socio-Ecological System

Without farmers there is no agriculture. Accordingly, a focus on pure yield effects and ecological aspects must remain incomplete if we want to find answers to the question of the development of the global food system. The effects on people in two of the most endangered regions are largely ignored: Sub-Saharan Africa and Southeast Asia. With a global warming of 3 degrees, agricultural labor capacity could drop by 30–50% due to heat stress. Food prices would rise, many more people would have to work in agriculture to maintain current levels. The global welfare loss at this level of warming could be as much as US$ 136 billion annually (de Lima et al., 2021). If one takes the total economic output into account, there would be a loss in gross world product of 23% (Burke et al., 2015, 2018). As a result of the development of regional heat centers, currently densely populated areas will no longer be suitable for agriculture and perhaps even become uninhabitable. This will lead to unpredictable migration movements (see also chapter “Escape from Heat, Drought, and Extreme Weather” Chazalnoel & Ionesco, (2024)), and even to the collapse of societies—on both sides of the migration movement (Kang & Eltahir, 2018).

Conclusion and Outlook

Under current climatic conditions, yield increases can be achieved by closing yield gaps. Only moderately intensified use allows yield increases of up to 80% on land that has been cultivated moderately intensively up to now (Beckmann et al., 2019). But as we have seen, these potentials are rather poor in a world that is 2–3 degrees warmer. Many theoretically achievable successes will be offset by expected negative side effects. However, it is also possible that the effects will negatively reinforce one another. Finally, different regions will be affected to very different degrees.

Against the backdrop of these uncertainties, simply tabulating future income losses and gains cannot be done in a reasonable way. The general trend, however, is clear and indisputable: a world that is 3 degrees warmer is confronted with the risk of massive yield losses; in extreme cases, large land areas might become desolate or uninhabitable.

Climate impacts affect agriculture all over the world: rich countries like Australia, rising nations like China (Kang & Eltahir, 2018), but also regions where a large part of the population must survive on less than $2 a day. There, the effects of climate-related yield reductions can be devastating, and the expected increase in extreme-temperature days can lead to major famines (Battisti & Naylor, 2009). Conflicts over resources are likely to intensify. Migration to the nearest metropolises or beyond will increase. We cannot rely on yields continuing to increase through intensification or expanded cultivation, or on global trade leading to a more equitable distribution of available resources. The gaps in supply which we haven’t been able to close in recent decades are likely to widen.

In a situation where we already produce twice as much food as we need, where approx. 40% is lost, most of it as food waste on, only 15% is due to harvest losses, a large part of the world’s population eats a diet far too rich in energy, we could today feed even a population of 10–11 billion. Thus, the solution does not call for more production. It is a collective task to mitigate climate change while ensuring agricultural yields under rapidly changing climatic conditions (Löwenstein, 2017). The solution rests on four pillars:

• 1. Yield security and stabilization with a high level of biodiversity must be ensured by means of agro-ecological principles and diverse crops of useful plants, whereby the latter must be adapted to future climates through innovative breeding;

• 2. Crop and food losses must be drastically reduced, ideally eliminated completely;

• 3. Eating habits must change toward a more conscious, lower-energy and healthier lifestyle;

• 4. Trade must serve to distribute food fairly and to compensate for possible climate-related yield losses, and must not displace locally adapted cultivation methods.

The question of how many people would be able to live and feed themselves healthily and sufficiently under the new climatic conditions of a 3-degree warmer planet cannot be answered with a sufficiently high degree of certainty. Of course, scenarios can be constructed, and well-researched contributions have been produced and is still worthwhile (ABC News, 2009). But it is not the scientific uncertainties of the individual effects compiled here in the chapter and the unclear effects of their combination that make an answer to this question seem speculative. More important is the uncertainty concerning our own decisions in the coming years: what resources we will consume in our societies, and in what quantities. It is our decisions that make forecasts uncertain but also provide some grounds for optimism.