image: Belief in the claim that irrigation agriculture produces 40% of all crops, as a function of time in polar coordinates. The size of the nodes (documents) correspond to their degree; that is, the number of incoming edges (citations). Blue nodes (“citation backup”) denote documents that make the claim and support the claim with a citation. Green nodes (“modeling”) represent documents that produce original data supporting the claim through a modeling or statistical exercise. Orange nodes (“no citation”) are documents that make the claim but do not produce original data nor cite any study to support the claim. Red nodes (“no claim”) are documents that are cited to support the claim but do not actually make the claim. Grey nodes (“NA”) represent documents that are cited as making the claim but that we have been unable to access.
Credit: Arnald Puy et al.
The idea that irrigated agriculture underpins global food and water security—producing 40% of crops and using 70% of freshwater—has become widespread in science and policy. However, these statistics are not empirically supported, according to a new analysis. Arnald Puy and colleagues traced these figures through citations in 3,693 scientific documents published from 1966 to 2024. The authors found that 60–80% of citation paths led to sources without supporting data or that did not contain the claimed numbers. Only approximately 1.5% of cited documents provided original data. When the authors analyzed available data on irrigation's actual impact, the results showed much wider uncertainty ranges: irrigated agriculture produces 18–50% of the world’s grain and is responsible for 45–90% of freshwater withdrawals. The analysis revealed that the 40% and 70% figures spread through “amplification," in which sources without data are used, and “transmutation," in which uncertain claims are presented as definitive facts after repeated citation. The most cited basis for the statistics, FAO's Aquastat database, contains limited country coverage with mostly imputed rather than measured values. According to the authors, the wobbly nature of such oft-repeated statistics highlights the need to critically evaluate foundational claims in sustainability science. Regardless of irrigation's global impact, there are no shortage of specific, local irrigation problems to solve, and they can often be approached most fruitfully with place-based solutions.
Journal
PNAS Nexus
Article Title
Widely cited global irrigation statistics lack empirical support
Article Publication Date
11-Nov-2025