Industry pollutes. This simple observation necessarily poses a problem for territories that intend to conduct a reindustrialization policy and preserve their biodiversity. As part of an internal exploratory study, INERIS approached this issue from the angle of accident risks, with three questions: 1/ If the main process accident scenarios (explosions, fires, toxic clouds) are taken into account in term of human consequences, one may wonder whether are not under-assessed / not assess at all regarding pollution. 2/ Following the logic of French methodological rules to produce achieve safety case, the hierarchy of ecosystem importance is a key parameter in the calculation of acceptability of a potential accident. The issue remains on the establishment of such Environmental scale: which ecosystem should be placed at the bottom of the importance scale, so that it is more permissible to pollute? Who makes such decisions today? On the basis of what knowledge? Within the framework of what legislation and regulatory practices? And what should be the right method and tools? 3/ Are we really prepared to deal with accidental pollution like Lubrizol in 2019 (www.aria.developpement-durable.gouv.fr/le-barpi/la-base-de-donnees-aria/) from a biodiversity perspective? Don't we tend to focus on prevention and, protection of people, without really knowing what to do to avoid and limit impacts on the ecosystems? Through an approach that combined an analysis of the literature on these issues (methods and orientations, theoretical reflections, accidentology) and a series of 21 interviews with key actors in France, the authors question a solution that seems better than the others, based on the notion of absorptive capacity/resilience of ecosystems. This approach would remain to be implemented for terrestrial ecosystems. This article presents the main steps of the demonstration.