Putting mathematics to work in the agricultural transition

The agricultural and agri-food sector remains one of the last sectors not using mathematical modeling to make decisions and manage risks. However, faced with climate challenges and the scarcity of fossil fuels, which require radical transformations in production models, this sector would benefit from adopting mathematics to better manage its transition. This would, however, imply a change in attitude towards nature and technology.

Mathematics, the great absentee from the agricultural sector

The agricultural sector is one of the last economic sectors where decision-making remains largely "empirical," based on experience and observation rather than theory or reasoning. Whether in the field of public policy, in cooperatives, or on farms, the use of theories or mathematical models to "rationalize" decisions, make them more objective, or even "enforceable" against third parties, remains rare. The agricultural sector favors the relational approach over the rational approach, that is, decision-making based on a consensus of experts or trusted peers guided by their experience and intuition, rather than on the results of mathematical or scientific model calculations.

This culture of empiricism is largely explained by the uncertainties and complexity specific to the agricultural context: open-air activity in diverse territories, hypersensitivity to an uncertain climate, living matter that is difficult to grasp clearly and, finally, a difficulty in formalizing expertise and practices that are often transmitted orally. Agriculture, a semi-industrial and semi-artisanal activity, is structurally difficult to rationalize, both in terms of vocabulary and processes[1], which leads its economic actors to adopt, for convenience, a more fatalistic than voluntary posture.

However, when we compare the agricultural sector with other economic sectors, we see that things could have been different, and that agriculture in the 1970s and 1980s truly missed the historic opportunity to use mathematics to modernize its decision-making. During this period, mathematical modeling experienced a meteoric rise in all economic sectors seeking to better control the complexity of their activity. Thanks to its ability to capitalize on expertise and model complex systems, so-called applied mathematics has established itself as an essential support for strategic and operational decision-making, from the banking sector to mass distribution, from transport to the manufacturing industry. Today, all these sectors use simulation, design, and management tools that rely on sometimes very sophisticated mathematical models.

If agriculture did not benefit from this phase of mathematization of its activity, it is because it chose to take a path radically opposed to that of other economic sectors: that of the simplification of its practices, relying on the two armed branches that were the mechanization of the countryside and the use of chemical inputs. This simplification approach has made it possible over the last 50 years to produce more food for an ever-growing world population and to domesticate the complexity of life with large reinforcements of fossil fuels, without the need to resort to mathematical modeling. But it has also had the consequence of weakening the agricultural sector by making it, in France, dependent on 98%[2] fossil fuels while impoverishing its traditional know-how.

Math to get out of fossils

The risks associated with global warming combined with the scarcity of fossil fuels make mathematics more necessary than ever today to guide the definition of public policies in all economic sectors, including the agricultural sector. In the energy sector, the study by RTE (Electricity Transmission Network), Energy Futures 2050[3] is a spectacular case study in this respect.

This study is based on a very comprehensive mathematical model that simulates the operation of the European electricity system every hour of every year for 30 years and integrates 200 weather reports from the Intergovernmental Panel on Climate Change (IPCC) tested at each of these hours. The RTE study thus makes it possible to objectively calculate 6 production scenarios and 3 energy consumption scenarios and to provide public and private decision-makers in the energy sector with rational elements to base their choices. In particular, it makes it possible to unambiguously arbitrate between the needs for energy from nuclear power and renewable energies by 2050 and to support, through mathematical results, the choice of French President Emmanuel Macron to launch the construction of 6 new EPR nuclear reactors (whether or not we agree with this choice).

The agricultural sector resembles the energy sector in many ways (fragmentation of production sites, uncertainties related to climate exposure, difficulties in storing production and globalization of trade associated with unpredictable price fluctuations). However, unlike the energy sector, in France, the primary sector does not have credible prospective studies based on mathematical models. Many studies exist, but they remain based on expert opinions, often carried out by associations with more activist than scientific vocations.[4] and none of them manage to reach consensus.

Change posture

The agricultural sector's 20-year delay in mathematical modeling is now a real handicap in its adaptation to ecological and climate challenges. Strategic decision-making leaves a lot of room for subjectivity, making consensus difficult and weakening public policies. The lack of mathematical risk modeling also slows the scaling up of transition financing by limiting it to voluntary and isolated initiatives. At the same time, the lack of mathematical modeling of the real impacts of food products is hampering the necessary reinvention of the economic model of agriculture and the remuneration of environmental services.

Like the hare in the fable, the agricultural sector would do well to catch up and invest massively in mathematical modeling before being forced to do so by worsening poverty and food crises. But to do so, it would first have to emerge from the slumber into which fossil fuels have placed it for the past 50 years and mourn the blessed time when food production was simple and predictable.

The agriculture of tomorrow will necessarily be much more complex and uncertain than that of the last 50 years. It will have to reconcile food production, environmental protection and the scarcity of fossil fuels, while ensuring decent incomes for farmers.[5]It will have to change its position towards nature, no longer considering it as a context to be domesticated but as a heritage to be maintained and protected.[6]Finally, it will also have to change its position with regard to mathematics and technology, using them both as a means of formalizing its practices and expertise and to model its trajectories in relation to the financial and agri-food sectors.

Driving the future

To seriously address the issue of food transition, public agricultural policies must now consider food chains as complex systems and high-risk activities, in which every link must be modeled, from farms to factories and all the way to consumers. They should draw inspiration from the way other critical supply networks for populations are modeled: water, energy, health, or public transport. They should work on the basis of iterative time-series scenarios, such as that of RTE, allowing for a fairly detailed understanding of the reality on the ground while assessing the impacts of decisions. Finally, it is necessary to integrate notions of margin of error, risk estimation, and crisis scenarios, which are currently completely absent.

This mathematization of agricultural policies is not a scientific fantasy but a moral necessity. To free ourselves from fossil fuels and have a chance of building a future for agriculture that is both efficient and sustainable for future generations, we must guide all our decisions in the short, medium, and long term with applied mathematics. This change of position will require a great deal of effort, as the agricultural sector has been, for decades, conditioned or even locked by the manna of cheap fossil fuels. Even if the old refrains inherited from the 1960s advocating the domestication of nature by fossil fuels resurface every time the specter of food insecurity resurfaces[7], we must now resist it courageously and invent the future of agriculture through a more mathematical and scientific approach than ever before.

Jeremiah Wainstain is the author of The food equation: feeding the world without oil by repairing nature and the climate, published by Editions La France Agricole in 2022.

[1] The many unsuccessful attempts to standardize an agricultural ontology or business process within ERP (Enterprise resource planning or integrated management software) are testimony to this.

[2] Souhil Harchaoui, Modeling transitions in agriculture: energy, nitrogen, and France's long-term food production capacity (1882-2016) and the beginnings of a global generalization, Doctoral thesis in Geography defended on December 9, 2019 at the University of Paris, p. 93.

[3] Energy Futures 2050: Production mix scenarios under study to achieve carbon neutrality by 2050, RTE. https://www.rte-france.com/analyses-tendances-et-prospectives/bilan-previsionnel-2050-futurs-energetiques

[4] See (non-exhaustive list) the work of ADEME, IDDRI, Solagro, WWF and the Shift Project in France.

[5] Interdependent objectives that are sometimes grouped under the vague concept of “regenerative” agriculture.

[6] A vision that environmental accounting and the CARE method in particular are trying to promote.

[7] The recent interview with Erik Fyrwald, head of the agrochemical group Syngenta, in the NZZ magazine is instructive in this respect. https://magazin.nzz.ch/nzz-am-sonntag/wirtschaft/syngenta-chef-fyrwald-fordert-ausstieg-aus-bio-landwirtschaft-ld.1683003