Environmental mitigation hierarchy and biodiversity offsets revisited through habitat connectivity modelling.

Biodiversity loss is accelerating because of unceasing human activity and land clearing for development projects (urbanisation, transport infrastructure, mining and quarrying …). Environmental policy-makers and managers in different countries worldwide have proposed the mitigation hierarchy to ensure the goal of "no net loss (NNL) of biodiversity" and have included this principle in environmental impact assessment processes. However, spatial configuration is hardly ever taken into account in the mitigation hierarchy even though it would greatly benefit from recent developments in habitat connectivity modelling incorporating landscape graphs. Meanwhile, national, European and international commitments have been made to maintain and restore the connectivity of natural habitats to face habitat loss and fragmentation. Our objective is to revisit the mitigation hierarchy and to suggest a methodological framework for evaluating the environmental impact of development projects, which includes a landscape connectivity perspective. We advocate the use of the landscape connectivity metric equivalent connectivity (EC), which is based on the original concept of "amount of reachable habitat". We also refine the three main levels of the mitigation hierarchy (impact avoidance, reduction and offset) by integrating a landscape connectivity aspect. We applied this landscape connectivity framework to a simple, virtual habitat network composed of 14 patches of varying sizes. The mitigation hierarchy was addressed through graph theory and EC and several scenarios of impact avoidance, reduction and compensation were tested. We present the benefits of a habitat connectivity framework for the mitigation hierarchy, provide practical recommendations to implement this framework and show its use in real case studies that had previously been restricted to one or two steps of the mitigation hierarchy. We insist on the benefits of a habitat connectivity framework for the mitigation hierarchy and for ecological equivalence assessment. In particular, we demonstrate why it is risky to use a standard offset ratio (the ratio between the amount of area negatively impacted and the compensation area) without performing a connectivity analysis that includes the landscape surrounding the zone impacted by the project. We also discuss the limitations of the framework and suggest potential improvements. Lastly, we raise concerns about the need to rethink the strategy for biodiversity protection. Given that wild areas and semi-natural habitats are becoming scarcer, in particular in industrialised countries, we are convinced that the real challenge is to quickly reconsider the current vision of "developing first, then assessing the ecological damage", and instead urgently adopt an upstream protection strategy that would identify and protect the land that must not be lost if we wish to maintain viable species populations and ecological corridors allowing them the mobility necessary to their survival.

Impact assessment of a high-speed railway line on species distribution: application to the European tree frog (Hyla arborea) in Franche-Comté.

The aim of the present work is to assess the potential long-distance effect of a high-speed railway line on the distribution of the European tree frog (Hyla arborea) in eastern France by combining graph-based analysis and species distribution models. This combination is a way to integrate patch-level connectivity metrics on different scales into a predictive model. The approach used is put in place before the construction of the infrastructure and allows areas potentially affected by isolation to be mapped. Through a diachronic analysis, comparing species distribution before and after the construction of the infrastructure, we identify changes in the probability of species presence and we determine the maximum distance of impact. The results show that the potential impact decreases with distance from the high-speed railway line and the largest disturbances occur within the first 500 m. Between 500 m and 3500 m, the infrastructure generates a moderate decrease in the probability of presence with maximum values close to -40%. Beyond 3500 m the average disturbance is less than -10%. The spatial extent of the impact is greater than the dispersal distance of the tree frog, confirming the assumption of the long-distance effect of the infrastructure. This predictive modelling approach appears to be a useful tool for environmental impact assessment and strategic environmental assessment. The results of the species distribution assessment may provide guidance for field surveys and support for conservation decisions by identifying the areas most affected.