Increasing ecological realism in conservation network design
a case study in Belize and an evaluation of global satellite telemetry for connectivity research
von Maarten Hofman
Datum der mündl. Prüfung:2017-05-15
Betreuer:Prof. Dr. Niko Balkenhol
Gutachter:Prof. Dr. Kerstin Wiegand
EnglischHuman-caused fragmentation and loss of natural habitat are among the world’s major challenges today. In combination with climate change, these processes contribute to soaring local and global extinction of many plant and animal species. Specifically, habitat loss and fragmentation disrupt contiguous natural areas, leading to low population sizes and geographic, demographic, and genetic isolation of populations. To avoid increased risk of extinction, conservation networks are created to preserve connectivity between the remaining patches of natural habitat. Oftentimes, conservation networks are clusters of protected areas connected through corridors or stepping-stone linkages. In applied conservation these networks have largely been based on expert-opinion. In landscape ecological research, more systematic modelling approaches are being developed to identify, evaluate, and optimise networks. These models aim to increase ecological realism in network design in order to avoid misguided management actions. However, increasing the ecological complexity in models requires more detailed ecological data, which is not always available in conservation settings. In this dissertation, I investigate the challenges and opportunities that arise when aiming for conservation network design based on models that use increasingly detailed ecological information. I used remotely-sensed landscape data and species detection data in Belize to model potential connectivity for white-lipped peccaries Tayassu pecari. This species is an endangered forest ungulate and acts as an umbrella species due to its large area requirements and relatively short dispersal distance. I included data on protected area effectiveness to determine habitat suitability, and estimated connections between high-suitability areas. I found that the model contributed to and augmented the current conservation network design by identifying alternative corridor routes and areas that were particularly important for connectivity conservation. Additionally, I deployed a satellite telemetry collar on a white-lipped peccary in southern Belize to obtain more detailed ecological data to parametrize the connectivity model. However, I found that data collection for the species was challenging, presumably due to the effects of forest cover and terrain ruggedness. I also observed that the success of GPS fixes was lowered by animal activity. The resulting data did not allow for connectivity modelling, but yielded an average home-range size estimate and relatively slow movement rates compared to other estimates for the species, and confirmed the species’ preference for forested habitats. Lastly, I conducted a global evaluation of satellite telemetry performance in wildlife research. I used a standardised questionnaire to avoid the bias towards successful implementations that is suspected to be present in the literature. I gathered information from over 3,000 telemetry units deployed on 63 species over 143 study areas, aiming to gain insight into the relative influence of the environment, topography, species characteristics, and unit specifications on the success of fix acquisition, data transfer and unit failure rate. I found that, in an average project, fix acquisition was relatively high but nonetheless only just satisfied researchers’ expectations. Species and unit characteristics were more important predictors of success rates than environmental factors. Data transfer rates were generally high, with satellite-based data transfer performing slightly worse. However, close to half of the deployments failed prematurely, and half of these suffered a technical malfunction. Understanding and modelling functional connectivity with increasing ecological realism is necessary for effective conservation network design. Network design based on moderately data-demanding models seems to be an achievable objective for current applied connectivity conservation initiatives. However, despite considerable developments in technology and analysis methods, modelling with high levels of ecological detail is still challenged by technological shortcomings and limited availability of detailed data. Ultimately, effective conservation network design depends on the continued collaboration between the modelling, empirical, and applied domains of connectivity conservation.
Keywords: white-lipped peccary; Belize; landscape connectivity; satellite telemetry; GPS-collar; movement ecology; habitat suitability; questionnaire; boosted beta regression; MaxEnt