Assessing the impacts of hydropower, dams and reservoirs on macroinvertebrate richness as a general portrait and within the Life Cycle Assessment (LCA) framework

Freshwater ecosystems are an essential resource to the survival of humans, namely through drinking and irrigation water, transportation and energy generation. The integrity and the function of these ecosystems strongly depend on their general state and biodiversity. Despite their dependance to these ecosystems, humans store freshwater into reservoirs and use it to produce energy via hydropower. This freshwater use leads to habitat degradation and destruction, and flow regime modification, amongst other threats. The scientific community has been studying the impacts of hydropower, dams and reservoirs on freshwater biodiversity for years now and a synthesis of these outcomes highlighted a wide range of impacts (both in their direction [positive or negative impact] and magnitude) dependent of multiple factors, such as taxonomic group studied, biomes and hydropower types, for instance.Life Cycle Assessment (LCA) is a methodology based on natural science that is used to evaluate the potential environmental impacts of a product, process or service, from resource extraction to end-of-life. Thus, to get an assessment that is representative of reality, we need to be able to account for the potential impacts of each phase in a product (or process, or service) life cycle. This includes the impacts associated with the energy source used to create these goods or services. Despite the mounting evidences regarding the impacts of hydropower, dams and reservoirs on freshwater biodiversity, LCA has only just started to account for these impacts into their assessment. Before that, impacts to freshwater biodiversity were not assessed and thus considered null, which is far from the reality. Accounting for theses impacts could influence the overall ranking of hydropower in terms of potential environmental impacts, when compared to other energy sources. Three issues led to the genesis of this doctoral research project; 1) in the review of the literature, hydropower, dams and reservoirs impact freshwater biodiversity but there is no consensus witHregards to the direction (i.e., positive or negative) and the magnitude (i.e., delta) of these impacts, 2) current Characterization Factors (CF) for the impacts of hydropower and reservoir occupation are exclusive to fish and/or based on theoretical curves instead of empirical data, and 3) CFs use Potentially Disappeared Fraction (PDF) of species (species level of taxonomic resolution), but there is still no study assessing the influence of taxonomic resolution (genus or family) on CFs. Thus, this project consists in three specific objectives:1. Qualifying and quantifying the impacts of hydropower, dams and reservoirs on freshwater biodiversity, using macroinvertebrate richness, through a thorough review of the literature and followed by a meta-analysis. 2. Developing regionalized CFs, using empirical data, to assess the impact of a riverbed transformed and occupied by a reservoir, hereafter referred to as the impact of reservoir occupation, on macroinvertebrate biodiversity (i.e., richness), within LCA. 3. Evaluating the influence of taxonomic resolution on previously developed CFs. For the first objective, we used a meta-analytic framework to uncover the overall direction and magnitude of the impacts of hydropower on freshwater macroinvertebrate richness. Our results showed a wide range of responses of macroinvertebrates to hydropower, dams and reservoirs, namely significantly negative, significantly positive and non-significant responses. Because of that large variability of responses, the meta-analysis did not show a significant overall impact of hydropower, dams and reservoirs on macroinvertebrate richness. Lots of environmental and methodological noise might have hindered the detection of a significant impact, which is why we also investigated the influence of some sources of heterogeneity and accounted for variables such as biomes, impact type, study design, as well as sampling seasons and gears. None of these variables showed a significant effect, which lead us to think that we may not have targeted the appropriate variables that could influence the richness of macroinvertebrates and that these organisms may be impacted by much finer-scale processes, such as granulometry, temperature, wave disturbance and macrophytes. For the second objective, we used a space-for-time substitution approach and developed empirical CFs (in PDF·m2·yr/m2·yr), to assess the impacts of reservoir occupation on macroinvertebrate richness at three spatial scales; reservoirs, ecoregions and the United States, as well as built an empirical model to explain the variability observed in reservoir PDF with only a few reservoirrelated environmental variables. In this objective, we observed a statistically significant loss of macroinvertebrate richness of 28% following reservoir occupation in the United States, that CFs specific to ecoregions followed a longitudinal gradient of impact, from lowest in the east to highest west, and that elevation, trophic state and reservoir surface area could be used to explain the variation observed in reservoir-specific CFs, where high elevation, oligotrophic and large reservoirs had higher CF. Our results provided strong support for regionalized CFs, showed that CFs were uniform and consistent across scales (i.e., reservoir, ecoregion and country scales) and offered a simple empirical model based on reservoir elevation, trophic state and surface area to be used by LCA modelers and practitioners to estimate CF where macroinvertebrate richness data is not available. For the third objective, which is considered as a complement to the second objective, we calculated CFs (in PDF·m2·yr/m2·yr) using richness data resolved at two levels of taxonomic resolution (i.e., genus- and family-based) and including a criterion of inclusiveness (i.e., exclusive and inclusive). Results in this objective suggest that using either genus- or family-based taxonomic resolutions conveys analogous messages. In both cases, regionalization of CFs is significant, directionality and magnitude of ecoregion CFs are very similar and empirical models are also similar with regards to model selection and predictive abilities. Based on these outcomes, we recommend favoring genusinclusive taxonomic resolution because they explained additional variation in CF and were more representative of the richness observed in the sampled ecosystems, as opposed to family-based and/or exclusive categories of taxonomic resolutions. The nature of this article is more fundamental and methodological. It has important implications for Life Cycle Impact Assessment (LCIA) practitioners because it could potentially allow the inclusion of certain groups of organisms, which are rarely identified up to the species level of taxonomic resolution, to be included in new CFs, thus enhancing the representativity and robustness of CFs for the ecosystems quality Area of Protection (AoP) in LCA. This thesis is very versatile and has many attributes. Some chapters are more fundamental, others have practical outputs. The choice of macroinvertebrate as research taxa is considered original, especially in the LCA community. Moreover, this thesis also showcases a novelty characteristic, especially the last article. To our knowledge, evaluating the influence of taxonomic resolution on CFs had never been done in the past. A major advantage of this thesis is its multi-disciplinarity, which makes its outputs especially relevant, deep and complex. It integrated biological data into an impact assessment tool aimed for decision making in the context of environmental sustainability of products, processes and services. The main contribution of this thesis is operational CFs assessing the impact of reservoir occupation on macroinvertebrate richness at three spatial scales, namely the United States, nine ecoregions and numerous reservoirs.