The Pacific Northwest climate varies along a latitudinal gradient, with increasingly longer, hotter and drier summers in the south. Recent experimental evidence suggests that warming in Pacific Northwest perennial grasslands favors invading annual grasses. Ecosystem responses to changing temperature, however, are likely to be non-linear and shaped by endogenous processes. For example, perennial grasses may tolerate some warming but be vulnerable to sustained or extreme periods of increased temperature. Once annual grasses establish, they may persist during cooler periods through feedbacks such as a thick leaf litter where their seeds can germinate, but seeds of other species cannot. If this is the case, we expect annual and perennial dominant states to exhibit a “hysteresis” dynamic where multiple ecosystem states are possible at the same environmental conditions, reinforced by internal feedbacks. Here we capitalize on the end of climate warming experiment to test barriers to perennial grass recovery. We hypothesize that annual grasses will maintain dominance post-warming due to litter feedbacks, but that burning will break this feedback and allow perennial recovery. We test this by factorially crossing burning (yes, no) by warming legacy (previously warmed, control) at three sites across a latitudinal gradient from southern Oregon to Washington. We add perennial seeds to each treatment and assess their survival.
Results/Conclusions
The warming experiment established a range of annual dominance as a baseline for perennial recovery with annual grasses invading southern and intermediate latitude sites. Plots became annual dominated in the south regardless of warming, but warming led to annual dominance at the intermediate site too. This suggests that warming encourages annual dominance, and southern ambient conditions are already in the range of hysteresis. When returned to ambient temperatures, plot composition remained largely unchanged in the first year, suggesting that annual grasses can persist after a heat wave. When we burned to remove litter, decreased annual cover suggests that litter does benefit annuals. We saw significantly higher establishment rates of seeded perennials in burned vs. unburned plots. Burning effectively removed litter, and litter levels were negatively correlated with perennial seedling survival. Our research shows that rapid, difficult to reverse state change may be imminent in Pacific Northwest grasslands, but suggests that burning followed by seeding perennials may be a useful management tool. We also provide evidence for a hysteresis dynamic in this ecosystem. Understanding how resilience will affect these dynamics will help managers prioritize conservation and restoration efforts.