Sustainability analysis and systems of linear equations in the era of data abundance

Sustainability research has entered an era of data abundance, in which online repositories offer millions of facts on production, consumption, pollution, and impacts. The combination of such facts in linear models leads increasingly to computational problems, relating to memory, speed, accuracy, and stability. This paper examines this phenomenon from the perspective of three widely-used types of sustainability analysis: multimedia fate and exposure models, life cycle assessment of products, and environmentally extended input-output analysis. The paper describes the various theoretical arguments, some indicators, and some solutions. Moreover, it adds the empirical evidence from one of these analysis types, namely life cycle assessment. It concludes that the phenomena indeed occur in practice, that abstract indicators ignore the subtle differences between different types of environmental impacts, and that a sound strategy for dealing with these problems is the critical analysis of results together with the variation of the computational principles.