Global water consumption impacts on riverine fish species richness in Life Cycle Assessment

Reduced river discharge and flow regulation are significant threats to freshwater biodiversity. An accurate representation of potential damage of water consumption on freshwater biodiversity is required to quantify and compare the environmental impacts of global value chains. The effect of discharge reduction on fish species richness was previously modeled in life cycle impact assessment, but models were limited by the restricted geographical scope of underlying species-discharge relationships and the small number of species data. Here, we propose a model based on a novel regionalized species-discharge relationship (SDR). Our SDR-based model covers 88 % of the global landmass (2320 river basins worldwide excluding deserts and permanently frozen areas) and is based on a global dataset of 11,450 riverine fish species, simulated river discharge, elevation, and climate zones. We performed 10-fold cross-validation to select the best set of predictors and validated the obtained SDRs based on observed discharge data. Our model performed better than previous SDRs employed in life cycle impact assessment (Kling-Gupta efficiency coefficient about 4 times larger). We provide both marginal and average models with their uncertainty ranges for assessing scenarios of small and large-scale water consumption, respectively, and include regional and global species loss. We conducted an illustrative case study to showcase the method's applicability and highlight the differences with the currently used approach. Our models are useful for supporting sustainable water consumption and riverine fish biodiversity conservation decisions. They enable a more specific, reliable, and complete impact assessment by differentiating impacts

Automated summary

This study proposes a model based on a novel regionalized species-discharge relationship to accurately assess the potential damage of water consumption on freshwater biodiversity. The model covers 88% of the global landmass and outperforms previous models used in life cycle impact assessment. It provides both marginal and average models with their uncertainty ranges and can support decision-making towards sustainable water consumption and riverine fish biodiversity conservation.