Illuminating Deep Uncertainties
in the Estimation of Global
Irrigation Water Withdrawals

Arnald Puy, Pamphile Roy and Andrea Saltelli propose a simple, easy to understand and computationally efficient approach for global sensitivity analysis based on the concept of discrepancy (the deviation of the distribution of points in a multi-dimensional space from the uniform distribution). The contribution provides modellers with a straightforward tool to conduct global sensitivity analysis.

Arnald Puy and Andrea Saltelli reflect on the philosophical foundations that ground the quest towards ever more detailed models, and identify four practical dangers derived from this pursuit: an explosion of the model’s uncertainty space, model black-boxing, computational exhaustion, and model attachment. The book is edited by Andrea Saltelli and Monica Di Fiore and is published by Oxford University Press.

In this Comment, Arnald Puy, Michela Massimi, Bruce Lankford and Andrea Saltelli discuss three epistemological obstacles that prevent irrigation models from providing accurate and "objective" estimates: the models’ elusive tie to reality, model plurality and indeterminacy of the target system.

Several modellers, irrigation scientists and experts in uncertainty show in Nature Communications that previous studies on global irrigation water withdrawals might have missed uncertainties spanning two orders of magnitude already at the grid cell level. Current estimates thus present us with a mirage of accuracy.