Modelling species distribution patterns using Posson Point Proceses

Specialeforsvar ved Neofytos Papapolydorou

Titel: Modelling species distribution patterns using Poisson Point Processes

Abstract: This thesis is devoted to spatial data modelling using the Poisson point process model while tackling the bias of the Citizen gathered data. Spatial data analysis is an emerging field of Geo-statistics and the increased computing capabilities make it possible to explore associations of spatial events and describe them through models. Spatial data differ from ordinary data in structure and spatial models are fitted using different features than standard regression models. The problem of bias in Citizen science is dealt with original techniques in similar papers and that is the approach followed in this thesis. Two methods are compared: Inverse Intensity Sampling is the main method and instrumental regression is the alternative. Both are novel techniques and there is no documentation that any of them is effective or not, at least in similar studies. For the examination of these approach-hes a dataset of fungal species recorded in Danish land cover in the last 15 years was used. The two methods enabled for a bias correction on the original model predictions.

Vejledere: Bo Markussen,   Jacob Heilmann-Clausen
Censor:      Søren Andersen, Novo Nordisk