Combining paleoecology and ecosystem modeling to study forest ecosystem consequences of wildfires from decades to millennia

Abstract:

Forest disturbances strongly regulate ecosystem structure and function, including vegetation composition, nutrient cycling, and energy flow. In ecosystems where disturbance is historically prevalent, vegetation and biogeochemical properties typically return to pre-disturbance conditions on time scales of years to decades. The biogeochemical impacts of changing disturbance regimes, however, are less well understood. Changing disturbance regimes may potentially impact ecosystem processes for centuries to millennia and alter global C and N balance. We combine a paleoecological record of fire and biogeochemical change with ecosystem modeling to test the hypothesis that variability in the timing and severity of wildfires has long-lasting impacts on biogeochemical processes in a subalpine forest in Rocky Mountain National Park, Colorado. We used an existing 4000-yr, high-resolution record of fire history (Fig. 1A) to drive the DayCent ecosystem model under two scenarios: (a) standard initialization where fire is repeated according to literature-estimated mean fire return intervals, and (b) paleo-informed scenario to dictate the timing of high-severity and non-high-severity fires. We compare the results to a suite of biogeochemical proxies from the same paleo record and draw inferences based on the two model scenarios.

When driven by the standard initialization, ecosystem properties of soil carbon, nitrogen, and net primary productivity (NPP) are relatively stable over the 4000-year period. In contrast, when driven with the paleo-informed disturbance history, soil C and N stocks and NPP undergo millennial-scale directional changes, similar in timing to some biogeochemical proxies from the paleo record (Fig. 1). In the model, these changes are due to a reduced frequency of high-severity fires after c. 2500 yr BP. Adding climate variability to this simulation is the next step for inferring potential mechanisms responsible for the millennial-scale trends observed in the paleo record. Our study demonstrates the relevance of centennial and millennial timescales for understanding modern biogeochemical states, which may prove critical for more accurately projecting future regional and global biogeochemical states through the use of ecosystem models.