Abstract
We used small-scale species distribution models to predict the past and present spatial distribution of 7 characteristic macrofauna species in response to climatic and environmental changes that have been recorded for the Jade Bay (German Wadden Sea) over the last 4 decades (1970s to 2009). Four presence-absence modelling algorithms (RF, MARS, GLM, GBM) were merged within the ensemble forecasting platform ‘biomod2’. The present spatial distribution (representing 2009) was modelled based on statistical relationships between species presences, true species absences and 7 high-resolution (5 m) environmental grids. The past spatial distribution (representing the 1970s) was then hindcast in response to climate change-induced (1) sea-level rise, (2) water temperature increase and (3) seagrass recovery due to de-eutrophication. The past distribution scenario was evaluated using independent historical macrofauna data from the 1970s. Present ensemble prediction maps accurately captured the potential ecological niches of the modelled species throughout Jade Bay (i.e. good to excellent true skill statistic [TSS] and area under the receiver operating characteristic curve [AUC] evaluation measures). The predicted present macrofauna distribution correlated most significantly with hydrodynamic conditions (submergence time, shear stress) and sediment characteristics (mud content). The past distribution scenario revealed significant changes in small-scale spatial distribution patterns of the characteristic modelled species (1970s to 2009) and showed a very good match with historical macrofauna data. Climate change-induced sea-level rise and its local implications for Jade Bay (changes in topography, tidal range and submergence time), and water temperature increase explained the potential macrofauna distribution shifts over the last 4 decades.
DOI Link to Publication