The new generation of individual-based forest models: what questions can be asked and answered

Background/Question/Methods

Individual-based forest simulators were employed in forest ecology since 1960s to forecast forest dynamics by simulating tree life cycle and spatial competition. Our ability to simulate dynamics of complex forested ecosystems is constantly improving in parallel with the development of new generations of computers making practical realization of computationally demanding algorithms feasible. Our conceptual understanding of forest dynamics is also substantially improving as forest inventory and acquisition of data via remote sensing come into being. Numerous modeling assumptions concerning both computational and biological complexities of previous models can be now substantially revised. This will improve the quality and predictive power of models and, accordingly, our understating of dynamics and complexity of forested ecosystems. In this presentation we delineate research questions that can be asked and answered with new generation models. We employ a new model called LES, which predecessors are ecological models of the JABOVA-FORET-SORTIE family and the forest management simulator, TACC. The novelties in the LES model extend the dimensions of the competition space. We reconsider crown competition for light, canopy vertical and horizontal structure and introduce belowground competition. A sunlight simulator predicts sun position depending on the geographical location, date and time, and computes light availability within the canopy.

Results/Conclusions

We have applied this model to address several questions that have not been accessible using previous generation models. Specifically: 1) We have discovered the connection between gap dynamics shade tolerance trade-off and water limitation, 2)  The model successfully simulated water-availability gradient  and corresponding transitions from grassland to forest, 3) We have simulated effects of different magnitude disturbances on species dynamics and composition, and stand macroscopic characteristics such as biomass, basal area, successional stage and diversity. In particular, we have quantified connections between the disturbance regimes and stand level changes and compered these results with forest transitions recorded in the FIA database, 4) We have also explored tree demography under different disturbance regimes and growth limitation by light and water availability, 5) The model allows us to predict spatial patterns observed when trees grow on the slopes and in the large latitude areas.  Overall, the new generation of individual-based models substantially expands our abilities to implement different functional trade-offs beyond classic shade-tolerance trade-off implemented in previous forest simulators.