Search

Understanding the impacts of changing climate and disturbance regimes on forest ecosystems is greatly aided by the use of process-based models. Such models simulate processes based on first principles of ecology, which requires parameterization. Parameterization is an important step in model development and application, defining the characteristics of trees and their responses to the environment, i.e., their traits. For species-specific models, parameterization is usually done at the level of individual species. Parameterization is indispensable for accurately modeling demographic processes, including growth, mortality, and regeneration of trees, along with their intra- and inter-specific interactions. As it is time-demanding to compile the parameters required to simulate forest ecosystems in complex models, simulations are often restricted to the most common tree species, genera, or plant-functional types. Yet, as tree species composition might change in the future, it is important to account for a broad range of species and their individual responses to drivers of change explicitly in simulations. Thus, species-specific parameterization is a critical task for making accurate projections about future forest trajectories, yet species parameters often remain poorly documented in simulation studies.We compiled and harmonized all existing tree species parameters available for the individual-based forest landscape and disturbance model (iLand). Since its first publication in 2012, iLand has been applied in 50 peer-reviewed publications across three continents throughout the Northern Hemisphere (i.e., Europe, North America, and Asia). The model operates at individual-tree level and simulates ecosystem processes at multiple spatial scales, making it a capable process-based model for studying forest change. However, the extensive number of processes and their interactions as well as the wide range of spatio-temporal scales considered in iLand require intensive parameterization, with tree species characterized by 66 unique parameters in the model. The database presented here includes parameters for 150 temperate and boreal tree species and provenances (i.e., regional variations). Excluding missing values, the database includes a total of 9,249 individual parameter entries. In addition, we provide parameters for the individual susceptibility of tree species to wind disturbance (five parameters) for a subset of 104 tree species and provenances (498 parameter entries). To guide further model parameterization efforts, we provide an estimate of uncertainty for each species based on how thoroughly simulations with the respective parameters were evaluated against independent data.Our dataset aids the future parameterization and application of iLand, and sets a new standard in documenting parameters used in process-based forest simulations. This dataset will support model application in previously unstudied areas and can facilitate the investigation of new tree species being introduced to well-studied systems (e.g., simulating assisted migration in the context of rapid climate change). Given that many process-based models rely on similar underlying processes our harmonized parameter set will be of relevance beyond the iLand community. Our work could catalyze further research into improving the parameterization of process-based forest models, increasing the robustness of projections of climate change impacts and adaptation strategies.

Abstract Understanding drivers of disturbances across scales is critical as environmental constraints change in a warming climate. Outbreaks of native bark beetles (Curculionidae: Scolytinae) are key natural disturbances that shape the structure and function of conifer forests across the northern hemisphere. While drivers of bark beetle outbreaks have been studied extensively at spatial scales ranging from stands to continents, within‐stand processes governing individual tree mortality in an outbreak are less well understood. Here, we use a spatially explicit long‐term monitoring dataset of a lodgepole pine (Pinus contorta var. latifolia) forest (>9000 individually mapped trees in three 2‐ha plots) impacted by a severe mountain pine beetle (Dendroctonus ponderosae) outbreak to explore interactions among fine scale drivers of beetle‐caused tree mortality. Using a Bayesian spatial modeling approach, we evaluated how tree scale and tree neighborhood scale characteristics interact with tree size to mediate host tree susceptibility to mountain pine beetle outbreak in the Southern Rocky Mountains (USA). We found evidence that both tree growth rate preceding the outbreak and neighborhood structure (within a 10 meter radius of the host tree) mediate the effect of tree size, and that the direction and magnitude of these mediating effects vary with tree size. Tree scale mortality probability increased with pre‐outbreak growth rate for small to medium sized host trees (~10–25 cm diameter), but that same effect was not detected for large trees. Conversely, tree scale mortality probability increased with greater neighborhood density, with the most pronounced effects for medium to large sized host trees (~15–30 cm diameter). Within‐stand topographic variability was not an important predictor of mortality probability; among stands, however, the stand in the driest topographic position experienced the greatest overall mortality. By explicitly considering how within‐stand heterogeneity mediates individual tree scale susceptibility to bark beetle outbreak, our findings bridge an important gap in understanding multi‐scale drivers of disturbance dynamics.