Table 1 Details of the simulation experiment used in this study. We simulated four factors on three landscapes over 400 years with five replicates and output generated every 10 years (400/10 × 5 = 200 observations per factor combination). No other disturbances but fire and MPB were simulated. We started all simulations using the initial conditions of the landscapes circa 2010. While the model generated over 33 output variables, we used only nine in this study. Variables with asterisks are the focus variables in this paper for brevity
Factor or variable | Levels, parameters, or units | Description |
|---|---|---|
Simulation design | ||
 Mountain pine beetles (MPB) | 3 | No beetles (none), mechanistic (Powell and Bentz 2014; Régnière and Bentz 2007), module (complex), and Amman/Keane empirical module (simple) (Keane et al. 2011) |
 Fire suppression level (FSL) | 2 | Historical wildfire regimes (no suppression), full suppression where 96% of fires are suppressed (suppression) |
 Fuel treatment effort (FTE) | 3 | No fuel treatments (none), and then a passive and aggressive treatment campaign where mechanical thinning of fire sensitive species and subsequent prescribed burns are done on approximately 2000 ha and 5000 ha per year, respectively |
 Climate (CLI) | 2 | Historical climate (HIST) with weather stations West Yellowstone (YCP), Prineville (OCH), and Boulder Mtn (ELK), and the hot, dry RCP8.5 IPCC (2011) scenario for the US Rocky Mountains (RCP8) where temperatures are around 5 °C warmer and precipitation is around 110% of historical climate: Offsets are Tmax 5.4, 5.2, and 5.5 °C and Tmin 6.5, 6.0, 5.5 °C for ELK, YCP, OCH, respectively, and precipitation 128%, 109%, and 102% more |
 Landscape (LAND) | 3 | Ochoco (OCH), Elkhorns (ELK), and Yellowstone Central Plateau (YCP). See Fig. 1 for map |
Simulation specifications | ||
 Simulation length | 400 years | Simulate landscape dynamics for 400 years to get a full representation of the fire and fire management regimes |
 Simulation replications | 5 reps | Needed to get sufficient observations to detect thresholds but low enough to generate enough observations for statistics without creating modeling bias |
 Simulation output interval | 10 years | Generated stand, site, and landscape level output at decade intervals |
 Time step | Daily | Cycle the 60–100 + year weather records for each of the landscapes for 400 years |
 Disturbance | None | No other insects/ disease impacts; no grazing; no human impacts |
 Total simulations | 540 | 3 MPB × 2 FSL × 3 FTE × 2 CLI × 3 LAND × 5 replicates = 540 |
Response variables | ||
 BAP* | (m2 ha−1) | Basal area of all pine species that are hosts to the MPB averaged across the landscape |
 HOSTAREA* | (%) | Percent landscape dominated by MPB host species |
 PLBURN* | (%) | Percent of the landscape burned annually |
 BA | (m2 ha−1) | Basal area of all trees averaged across the landscape |
 CBD | (kg m−3) | Canopy bulk density |
 CWD | (kg m−2) | Coarse woody debris-downed woody material greater than 10 cm diameter |
 FWD | (kg m−2) | Fine woody debris- downed woody material less than 10 cm diameter |
 BAMPB | (m2 ha−1) | Basal area of all trees killed by MPB |
 PLFAS | (%) | Percent of the landscape composed of fire adapted species |