Druckansicht der Internetadresse:

Lehrstuhl für Biogeografie

Prof. Dr. Carl Beierkuhnlein

Seite drucken
Jaeschke, A*; Bittner, T; Reineking, B; Beierkuhnlein, C: Modelling biotic interactions facing climate change: a comparative study of three different approaches
Vortrag, Jahrestreffen des AK Biogeographie im Verband der Geographen an Deutschen Hochschulen (VGDH), Hamburg: 20.05.2011 - 22.05.2011

Abstract:
Climate change is expected to alter biotic interactions leading to temporal and spatial mismatches of interacting species. However, the implementation of biotic interactions in species distribution modelling is rarely taken into account and poses some challenges. Here, we examined potential impacts of climate change on the biotic interaction between the Green Hawker (Aeshna viridis) and its only egg-laying plant Water Soldier (Stratiotes aloides) in Europe using an ensemble forecasting modelling technique (BIOMOD). We compared three different approaches of considering this biotic interaction in species distribution modelling: (1) an ‘overlap approach’, modelling the potential future distributions of both species individually and intersecting the range overlap, (2) an ‘explanatory variable approach’, modelling the potential future distribution of A. viridis with the projected current distribution of S. aloides as additional explanatory variable beside climate, and (3) a ‘reference area approach’ modelling the current distribution of A. viridis in the current range of S. aloides (model fitting) and projecting its potential future distribution on Europe. We present the modelling results and a comparative analysis of approach performance. According to the models, the biotic interaction between these two species tends to remain stable for all approaches. Nevertheless, the approaches differ in their predictive performance related to the AUC and in their span width concerning the projected range change between the different modelling algorithms. The question of how to integrate biotic interactions in species distribution modelling is not easy to answer. In this case, all approaches performed well and produced reasonable results.
TwitterInstagramYoutube-KanalKontakt aufnehmen
Diese Webseite verwendet Cookies. weitere Informationen