The objective of this project was to use an ecosystem model to simulate the effects of management actions and disturbance on forest carbon dynamics. The central question was: What are the carbon tradeoffs of different management actions designed to meet different management objectives? Different management scenarios were evaluated at three different installations (Fort Benning, Camp Navajo, and Joint Base Lewis McChord), each with different management histories and a focal species of wildlife with specific habitat requirements.

The project began with the intent of coupling an ecosystem model with a growth-and-yield model by using data on whole tree leaf biomass. However, for reasons including model performance and feasibility of technology transfer to installation managers, the project team transitioned to using a different modeling approach. The landscape-scale succession and disturbance model, LANDIS-II, was used because it performed well in validation and presents fewer challenges for technology transfer. The model uses an age-cohort based approach to simulate forest succession and growth over time and across space. In addition to the core model, the Century Succession extension was used, which enables the simulation of aboveground and belowground carbon stocks and fluxes. The model framework was parameterized for each of the installations and used to simulate carbon dynamics as a function of different management scenarios.

Increasing treatment intensity resulted in decreasing total ecosystem carbon. Treatments that included thinning, however, increased net ecosystem carbon balance, making thinned forests a larger sink for carbon. This finding is a result of decreased competition for resources (e.g., water, nutrients, light) resulting from thinning. At Fort Benning, actively restoring longleaf pine forest decreased total ecosystem carbon by approximately 22% compared to the control (fire-suppressed broadleaved forest). Longleaf pine (Pinus palustris) restoration, however, increased red-cockaded woodpecker (Picoides borealis) habitat such that approximately 90% of the upland forest area (historically longleaf pine dominated forest) was viable habitat by the end of the simulation period. At Camp Navajo, thinning and prescribed burning treatments decreased total ecosystem carbon relative to the control in the absence of wildfire. When wildfire was included in the simulations, the thin and burn treatment had higher total ecosystem carbon than the control by the end of the simulation period. This results from the reduced risk of high-severity wildfire from treatment. Simulations also demonstrated that thinning and burning decreases the risk of habitat loss for the Mexican spotted owl (Strix occidentalis). At Joint Base Lewis McChord, treatments that included thinning resulted in the lowest total ecosystem carbon, but the highest net ecosystem carbon balance. In addition, thin-only and thin and burn treatments significantly increased the probability of Oregon white oak (Quercus garryana) presence by the end of the simulation period. This species provides an important food source for the western gray squirrel (Sciurus griseus) during mast years.

Longleaf pine forest (Pinus palustris, left), ponderosa pine forest (Pinus ponderosa, middle), Douglas-fir forest (Pseudotsuga menziesii, right).

Workshops were held at each of the installations to train resource managers in the operation of the parameterized models. The models provide a planning tool for evaluating the cumulative effects of stand-scale management actions on installation-wide forest carbon stocks and sequestration.