Skip navigation
Previous Next

Science News

February 27, 2014




Emiel van Loon

University of Amsterdam


Friday 28 February 11:00

Sir Neil Chalmers seminar room, Darwin Centre LG16 (below Attenborough studio)


Species distribution modelling (SDM) is increasingly applied to answer all sorts of ecological questions. Implicitly, the
SDM literature suggests that the question of interest together with the available data prescribe the appropriate methods for data analysis. In contrast with this suggestion, it can be argued that a range of questions concerning the distribution of a species is usually interesting, while the available species occurrence and environmental data cannot easily be changed. Hence it may be effective to establish which questions may be answered by the available data as a first research step. The number of species records is one of the most important factors limiting the research questions and methods that are applicable. For that reason this presentation will focus on the relation between the number of available species records and the potential to answer different research questions.


First a hierarchy is proposed to organise research questions that differ in nature and complexity, and to cast different research questions in a model comparison framework. Using this framework, research questions of different complexity are translated into SDMs. Through different simulation examples, the effect of the number of occurrence data on the possibilities to identify SDMs with different numbers of predictor variables as well as on predictive performance are shown. Next, it is shown that with increasing scarcity of species records, either the information requirement as dictated by the research questions has to decrease or more prior knowledge about the species-environment or geographical relations have to be specified. The presentation concludes with a preliminary overview of research questions on species distributions and the matching levels of occurrence records that are required to obtain an adequate answer.


For additional details on attending this or other seminars see