dc.description.abstract
The declaration of protected areas (PAs) is generally considered one of the main tools to counteract the constant decline in global biodiversity. Despite a strong increase in global PA-coverage during the last two decades, the downwards trend in biodiversity has not yet stopped or even reversed. Freshwater ecosystems, both lotic and lentic, are especially species rich, but show stronger biodiversity declines compared to terrestrial or marine ecosystems. Unfortunately, freshwaters are often not explicitly considered in the declaration of PAs and, as one consequence, freshwater biodiversity is underrepresented in existing PA-networks. Future climate change is predicted to further impact the effectiveness of existing PAs, because of a decrease in habitat suitability within PAs for target species. Hence, there is an urgent need to anticipate the impacts of climate change and other emerging pressures to adjust existing, or declare new PAs to optimise them for the future. In addition, the growing demands of the human population for space and natural resources aggravate the complexity of nature conservation efforts.
The aim of my thesis was to close some of the related knowledge gaps to protect freshwater biodiversity under future climate change. My research showcases, with the example of the upper Danube river basin, how predictive models can be applied to river ecosystems and how model outputs can be used to identify future possible environmental pressures, their temporal dynamics, and how conservation management can be informed by this knowledge.
The thesis is divided into three parts. First, I investigated how spatial resolution affects model outcomes in river habitat suitability models (HSMs). Second, I assessed if environmental pressures for riverine fish species differ between the past 200 years and in the future until 2100, by using climate niche factor analysis. Third, I analysed if the established network of PAs within a selected study area protects native fish species under current environmental and future climate change conditions. Third, I used these findings to optimise the existing network by adding individual PAs based on a systematic conservation planning approach, aiming at sufficiently protecting native fish species and selecting PAs that will serve as environmental refugia in the future.
Based on a compiled extensive fish occurrence database, I used ensemble HSMs to relate the occurrence of 48 native fish species with environmental parameters including topography, land-use, climate and hydrology at the respective locations. HSMs were calculated on a sub-basin level as modelling unit and for ten different spatial resolutions (i.e. average sub-basin size), while keeping all other model parameters constant. With this approach, I showed that predictor importance (which is a measure of which model parameters are more or less important to predict fish habitat) and predicted suitability patterns (i.e. the distribution of areas that are predicted to be suitable as fish habitat) are highly dependent on the spatial resolution of the model. Furthermore, by correlating predicted suitability among nested sub-basins, I was able to identify a scale tipping-point at which the set of environmental parameters predicted habitat suitability patterns best (i.e. habitat suitability patterns did not improve even if a finer scale was applied).
I used the species-specific distribution patterns at the finest modelled resolution together with a unique time series of modelled and observed hydrological and climate data for 1800–2100 to analyse how environmental pressures for riverine fish species differ in type, spatial distribution or both between the past 200 years and in the near future in 2100. I used climate-niche factor analysis to calculate species-specific vulnerabilities, i.e. the magnitude of a species to be impacted in a specific location, and the driving environmental pressures. I showed that historical and future environmental pressures resulted in similar vulnerability estimates for native fish species which were, however, caused by different environmental pressures in 1800 compared to 2100. Historically, fish species were mainly impacted by a change in hydrology (more specifically by a decrease in the variance of monthly discharge), while in the future temperature will be the main pressure (i.e. the predicted increase in mean annual temperature). This change in main environmental pressures was accompanied by a spatial shift of areas that were predicted to be especially impacted.
To investigate whether native fish species are sufficiently protected within the existing network of PAs, currently and in the future, I grouped all native fish species according to their threat status based on the IUCN Red List categories. Then, I analysed the coverage of each group within the PAs and their changes in vulnerabilities under future climate change scenarios within the existing PA-network. I found that the existing PA-network currently insufficiently protects the distribution range of native fish species when applying the 20-60% guidelines (i.e. protect 20 -60% of a species distribution range depending on the species ecology, distribution, and population trends) suggested by the European Commission. Consequently, the conservation planning analysis revealed that an additional c. 6000 sub-basins need to be added to the c. 2000 current sub-basins to reach sufficient protection for native fish species. In addition, I showed that the existing PA-network is located in areas in which fish species will be especially exposed to future pressures from climate change. To spatially optimise the PA-network under future climate change conditions, I used the magnitude of predicted climate change-induced habitat alterations in each sub-basin as a cost factor in the systematic conservation planning analysis. Therefore, the proposed PA-network is located in areas that experience the lowest habitat alterations in the future. For all planned networks (current and future) I found a high spatial overlap, indicating that a currently optimised network can also safeguard native fish species under future environmental conditions.
With this thesis I demonstrated the feasibility of predictive models to identify habitat suitability patterns of native fish species in river ecosystems. In addition, I showed how model results can be used to identify areas that are predicted to especially suffer climate change impacts and how such predictions can inform conservation planning analysis. The results of the conservation planning analysis can directly inform the revision of the existing PA-network in the Upper Danube River basin to effectively protect native fish species into the future.
en
