CoFoLaMo: Comparison of Forest Landscape Models

Forest landscape models are used to simulate changes in forest structure and composition at the landscape scale (~10²-10⁶ ha) over decades to centuries. There are several landscape models, and the basic structure of these models is similar. They divide the landscape into grid cells with heterogeneous environmental conditions and simulate the demographic processes (growth, regeneration, mortality) within these cells, as well as spatial interactions between these cells (e.g., seed dispersal and spatially explicit forest disturbances such as bark beetle infestations or forest fires). However, despite these similarities, forest landscape models differ considerably, for instance with regard to their spatial and temporal resolution or the formulation of processes and spatial interactions. Due to complex feedback mechanisms, even comparably minor differences may lead to noticeable between-model differences in future projections. To better assess the predictive power of such models, it is important to quantify the uncertainty that arises from different model formulations. However, a quantitative comparative study is missing for forest landscape models.

In this project, we conduct the first systematic comparative study of forest landscape models. The following models are included: LandClim, TreeMig, LANDIS II and iLand. The first objective is to quantify the uncertainty in future projections. For this purpose, we carry out a large number of simulations under different climate, disturbance, and dispersal scenarios using harmonized model input data, analyze differences in simulation results, and quantify the proportion of variance explained by model differences. The second objective is to identify the main causes for observed differences between models. To this end, we systematically compare model structures, formulate hypotheses about probable causes of simulated differences, and conduct simulation experiments to test these hypotheses. This project is an important steps step towards more robust forest landscape models.

Funding

This project is part of the COST Action external pageFP1304call_made: Towards robust projections of European forests under climate change (PROFOUND).

Status of the project

This project runs from September 2016 to March 2019.

Contact

Please contact Gunnar Petter for more information.