PS 32-195 - Initialization of LANDIS-II with TreeMap

Abstract: Forest succession, or ecosystem process models, can aid in many land management decisions. They inherently incorporate ecological processes (competition for nutrients, water, and light; dispersal) to predict how additional disturbances (wild and prescribed fire, climate change, insect outbreaks, harvest) impact the resilience of the forest and its potential future structure (species types, ages, biomass; above and belowground carbon storage and sequestration potential). Many land managers need projections on landscape change and carbon storage (long term) and prescribed fire implementation (short and long term). The landscape class of ecosystem process models are often ideal for these goals as they simulate individual tree species interactions through time and offer a robust assortment of disturbance extensions. However, broad implementation and adoption are often limited as users must rebuild the landscape models (soil and climate properties; initial tree communities – type, age, amount) for every new site of interest. To address this issue, we constructed a common parameter set for the landscape class ecosystem process model, LANDIS-II, for the entire USFS Region 8 (the 13 southern states). Forested communities within the model were constructed using TreeMap. TreeMap is a gridded CONUS dataset that imputed Forest Inventory and Analysis (FIA) plots based on landscape data (topographic, biophysical, and disturbance) to provide wall-to-wall (30m x 30m) predictions of which FIA plot best represents any given forested pixel. Total species were reduced based on FIA species group code (~30 eastern codes) and the characteristics necessary for LANDIS-II simulation (ex: age and biomass of species) calculated. We collated the necessary tree species and soil parameters for all sites within the study area and provided a climate driver for simulation. The final tool allows users to import a shapefile and extract all necessary model inputs for that region. These inputs allow for rapid initial analysis without the user needing the data assimilation techniques commonly used in these projects. Further, these methods are robust and flexible enough to expand this work to CONUS and incorporate additional species and other LANDIS-II extensions. In conclusion, this methodology and tool can serve as a starting point for rapid initial analysis of many land management and research questions.