University of Twente Student Theses


Comparison of autecological and synecological approaches in modelling present and future climate distribution of rare and endemic alpine plant species in Lefka Ori (Samaria) National Park- Biosphere Reserve, Crete

Iseghohi, Dorothy Agbonghale (2021) Comparison of autecological and synecological approaches in modelling present and future climate distribution of rare and endemic alpine plant species in Lefka Ori (Samaria) National Park- Biosphere Reserve, Crete.

[img] PDF
Link to full-text:
(only accessible for UT students and staff)
Abstract:Anthropogenic warming is anticipated to impact and change species communities globally. The Mediterranean, a hotspot of endemics, is projected by many global and regional estimates of future climate scenarios to have less precipitation and rising temperatures, thus more vulnerable to climate change. As a result, the impacts of climate change are anticipated to be immense on Mediterranean islands and mountains. Plant species within this alpine region will be vulnerable to climate change, and a potential loss of suitable habitats is expected in the future. Several studies have extensively used species distribution models (SDM) to predict and model the impacts of climate change on species. These SDMs are commonly used to model individual species and make predictions of plant distribution in relation to the environment in an autecological way. SDMs are not built to account for species associations and communities of species in a single model. Nevertheless, species are associated, and some form of biotic interactions affects species co-occurrences within plant communities and the spatial distribution of plants. Hence, Joint Species Distribution models (JSDM), which is an improvement of the common SDMs, are structured to capture biotic interactions between species in relation to the environment and predict the distribution of plants in a Synecological way. JSDMs offers the advantage of modelling species communities in a single model and depicts species co-occurrences. The alpine plant species on Lefka Ori Mountains in the Island of Crete, Greece, is the focus of this study. It was aimed to compare JSDMs with SDMs in predictions of alpine plant species distribution and determine whether a predicted distribution range shift resulting from climate change is different in JSDMs compared to SDMs. The alpine species dataset used in this study comprises forty-two rare and endemics, and environmental variables include altitude, curvature, distance to road, geology, snow, aspect. The Hierarchical Modelling of Species Community (HMSC) is used to model the 42 plant species in a single model (JSDM). Furthermore, five individual species were selected through a residual correlation matrix derived in JSDM and other parameters. They were modelled on their own with the same HMSC as SDMs and compared with JSDMs. Predictions from both models are extrapolated using climate proxies of snow and altitude to model the climate change scenario by a century to derive potential species range shifts and suitable habitats shifts. JSDM and SDMs were compared through prediction accuracies, regression coefficient matrix, variance proportions, range shift, and suitable habitats derived from climate change scenarios. This study found that JSDM models the potential biotic interaction between species, resulting in identified co-occurrences within the species community. The study's statistical analysis revealed a significant difference in the mean AUC values of JSDMs and SDMs. In addition, the accuracies of co-occurring species are lower in JSDMs than SDMs, while the non-co-occurring species had lower accuracies in JSDMs. The important variables contributing to plant distribution are geology, snow, curvature, road and altitude. The proportion of variances are slightly higher in SDMs than JSDMs. The results from the future climate scenario indicate suitable areas net lost, and these losses are more pronounced in JSDM than SDM for all five species investigated. Lastly, the results of this study indicate that the potential future changes to species suitable habitat due to climate change are species-specific and not dependent on the modelling approach
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page