Impact of Species Parameter Uncertainty in Simulations of Tree Species Migration with a Spatially Linked Dynamic Model

The simulation of tree species migration suffers from many sources of uncertainty. In our study we examined the influence of species parameter uncertainty on simulated tree species migration, using the spatially linked dynamic forest landscape model TreeMig. The impact of uncertainty becomes especially apparent under critical conditions arising from the interaction of species sensitivities to climate, species competition and spatial fragmentation. Therefore we examined the differences in migration success and speed in a realistic scenario including these critical conditions. The south-north migration of the submediterranean tree species Ostrya carpinifolia through the highly fragmented and climatically heterogeneous landscape of the Swiss Alps was simulated for 27 different species parameter sets covering the plausible range of species parameters for O. carpinifolia. To account for the additional uncertainty introduced by the stochastic representation of future climate variability, each species parameter set was simulated with multiple repetitions. We found that migration success and speed resulting from simulations with the different sets varied highly. The current situation of rapid climate change and high landscape fragmentation due to human land use could create critical conditions comparable to the simulated scenario for various species. We therefore recommend testing for species parameter sensitivities and – if indicated – to repeat simulations with different parameter sets when projecting future tree species distributions with explicit simulation of migration.