LEE- Yer Sistem Bilimi-Doktora

Bu koleksiyon için kalıcı URI

http://hdl.handle.net/11527/19501

Gözat

Konu "Ajan tabanlı modelleme" ile LEE- Yer Sistem Bilimi-Doktora'a göz atma
  • Öge
    Invasive species distribution modeling under climate change
    (Graduate School, 2023-08-08) Kanmaz, Oğuzhan ; Dalfes, Hasan Nüzhet ; 601152001 ; Earth System Science
    Biological invasions are one of the most significant components of global environmental change. The biological consequences of invasions such as biodiversity loss, biotic homogenization, alterations in the structure and function of ecosystems are difficult to reverse and even impossible in many cases. The damages and loss caused by biological invasions are estimated to reach billion dollars each year. Post introduction control of an invasive species is often futile alongside being expensive and labor intensive. In this sense, prevention is considered to be the most effective strategy against invasions. Determination of the areas under risk of invasion and evaluation of the potential invasion scenarios are extremely important. To that end computational models constitute crucial tools. In recent decades increased availability of powerful hardwares, alongside with the accessibility of environmental and biogeographic data, due to the developments in information technologies, lead to the utilization of computational models in various fields of ecology. Species Distribution Models (SDMs), which are an example of such applications, especially the data-driven correlative methods, were utilized widely to investigate the impacts of global environmental change on the current and future distribution of species. Application of SDMs on invasive species to determine the areas under invasion via the projected suitability/presence probabilities, despite the criticism on the violation of the equilibrium assumption, has become a widely used method with various successful and promising examples. However, another problematic aspect of such applications is that the projected suitability cannot always be interpreted as an actual invasion since the dispersal is a crucial process of biological invasions. Agent-Based Modeling (ABM) is a population modeling method with wide applications in ecology to simulate various complex processes via the functional units called agents which can interact with each other and the environment. In this respect agent-based models constitute a remarkable alternative to simulate biological invasions with various examples in the literature. While agents-based models are especially useful to investigate the theoretical problems and considered as in silico laboratories for paradigmatic models, their application on spatially explicit, real world cases to construct pragmatic models is mostly difficult due to the lack of a priori knowledge on the majority of the species to calibrate such models. Hybrid modeling is a promising approach which can utilize more than one modeling method in tandem, by taking advantageous aspects of these methods which can complement each other to obtain more reliable results. Within the scope of this thesis, a hybrid modeling framework, which consists of a correlative and an agent- based component was constructed. Accordingly, while the framework proceeds in yearly time steps, correlative component produces bioclimatic suitability projections to be used by the agent-based component and the agent-based component generates simulated occurrence records to be utilized by the correlative component for making projections. Beyond the basic structure of the framework, the modular structure enables the inclusion of species specific processes which are constructed based on the a priori knowledge to obtain more realistic projections. For the implementation of the modeling framework, Impatiens glandulifera, a highly aggressive invasive plant native to Himalayas, was selected. Since its introduction to Europe and North America in the late 19th century as an ornamental plant, it has primarily invaded riparian habitats. The current invasive range of I. glandulifera spreads across the northern hemisphere. In the last two decades, it was observed to invade forests and mountainous areas. The spatial context of the model was determined as North America where the invasive range is far from reaching its bioclimatic potential and the simulations were conducted for 2020-2050 period under RCP 4.5 climate change scenario. In accordance with the a priori knowledge on I. glandulifera, the agent-based component of the framework which consists of three procedures (Climatic Window Procedure, Landscape Suitability Procedure and Propagule Procedure) to process three types of agents (productive, post-production and pre-productive agents) was constructed. The Climatic Window Procedure which consists of Chilling Period, Bioclimatic Suitability, and Productive Agent Sampling sub-procedures, performs the transformation of pre-productive agents to productive agents based on the chilling requirement and bioclimatic suitability projections that are generated by the correlative component alongside the sampling of the generated agents. The Propagule Procedure performs the generation and dispersal of post-generation agents. The Landscape Suitability procedure evaluates the transformation of post- generation agents to pre-productive agents based on the pH, elevation, slope, and land use properties. As the result of the conducted simulations, it was observed that the initial invasive range in North America, which is on the west and the east shores initially, has expanded through the 2020-2050 period. While in the eastern part of the invasion range, Great Lakes region and New England shores were observed to be saturated, progression on the western part was primarily determined by the mountain ranges. The severity of the projected invasion range on the Alberta-Saskatchewan region is especially remarkable, considering the limited presence of I. glandulifera in this area in the initial conditions. Another important result is the potential formation of a continuous range crossing the continent in the long term, due to the aggressive expansion to the interior regions. The projected latitudinal progression of the invasion is compatible with the large- scale pattern of the northward progression of the species due to climate change. While the northern boundary of the projected invasion range was roughly following the boreal biome, the southern boundary was formed on latitudes similar to the southern boundary of the invasion range in Europe. Another striking similarity was the southern progression on the west shores being limited by the Mediterranean climate observed in the region, which is also the determinant of the limited southward progression in Europe. The potential impacts of the I. glandulifera on the boreal forests, which are expected to be more fragile in the future due to climate change, must be carefully evaluated since I. glandulifera is known to spread in such disturbed forests. The long distance dispersal is mostly a result of anthropogenic activity, and unlike many of the natural processes, can lead to unpredictable results. This constitutes a limitation for the projections. Also, the impact of the incomplete occurrences data on the projections, especially records for the regions, is unavoidable. In the scope of this PhD thesis, it was aimed to construct a modeling framework by the utilization of correlative species distribution and agent-based modeling methods in tandem to simulate biological invasions under climate change. The framework was implemented to make projections for I. glandulifera invasion in North America for 2020-2050 period under RCP 4.5 scenario. In accordance with the conducted spatiotemporal analysis on the projections, invasion patterns were determined and the potential impacts on the invasive range were evaluated. The results were observed to be in concert with the historical invasion patterns of I. glandulifera in Europe and the climatic and environmental projections for North America. In this regard the framework can be considered a promising tool to be utilized for making projections which can be used for determination of the areas under invasion risk.