Repository logo

Forest recovery trends following an unprecedented mountain pine beetle outbreak in the southern Rocky Mountains




Wykhuis, Aaron, author
Ex, Seth, advisor
Fornwalt, Paula, advisor
Sibold, Jason, committee member

Journal Title

Journal ISSN

Volume Title


Following the mountain pine beetle (MPB) outbreak in the early 2000s in the southern Rocky Mountains, questions remain about what the structure and composition of lodgepole pine forests will look like as they recover. I used Forest Inventory and Analysis data collected by the US Department of Agriculture Forest Service, together with hierarchical cluster analysis, to separate 626 subplots into unique forest type groups based on their post-outbreak overstory and understory stand structure and composition. The subplots had all been dominated by lodgepole pine in the overstory prior to the outbreak and had all experienced at least 10% basal area mortality. I verified the forest type groups through multi-response permutation procedures and characterized them through indicator species analysis. I used a random forest analysis to determine how a suite of abiotic and biotic variables were predictive of the forest type groups. Post-outbreak stands separated into five forest type groups. These groups all remained dominated by lodgepole pine in the overstory but had differing understory structures and compositions. Naming conventions for the groups were based on their "overstory/understory" dominance. The five forest type groups were lodgepole pine/lodgepole pine saplings (PICO/PICO.saps), lodgepole pine-subalpine fir/subalpine fir (PICO-ABLA/ABLA), lodgepole pine/sparse understory (PICO/Sparse), lodgepole pine/lodgepole pine seedlings (PICO/PICO.seeds), and lodgepole pine/quaking aspen (PICO/POTR). The six variables with the highest relative importance in predicting forest type were (in order) overstory basal area mortality, precipitation as snow, maximum July temperature, pre-outbreak basal area, pre-outbreak stand age, and precipitation as rain. Each of the five forest type groups varied in their probability of occurrence along gradients of the top six predictor variables. For example, the PICO/PICO.saps and PICO/Sparse groups were most likely to occur where basal area mortality was low, the PICO-ABLA/ABLA group was most likely to occur where basal area mortality was moderate to high, and the PICO/PICO.seeds and PICO/POTR groups were most likely to occur where basal area mortality was very high. Overall, my results suggest that as southern Rocky Mountain lodgepole pine forests affected by the MPB outbreak recover in the near-term, they will remain dominated by this species. However, they will have more structural and compositional diversity owing to a high degree of heterogeneity in abiotic and biotic factors, which will likely shape longer-term outcomes.


Rights Access



Associated Publications