Vegetation Feedback Mediates Climate Warming Effects On Insect Disturbances In A Temperate Deciduous Forest

Abstract:

Climate warming directly affects insect disturbance regimes by altering temperature-dependent population development. Successional changes driven by climate warming in host tree demography, which is referred to as vegetation feedback, can also affect insect disturbance regimes. However, such feedbacks have rarely been incorporated in insect disturbance regime predictions. Moreover, few studies have predicted insect disturbance beyond the 21st century and thus, it is unclear how insect disturbance will be affected by climate change in the long-term. We implemented insect development and vegetation feedback into a forest landscape modeling framework to simulate red oak borer (ROB, Enaphalodes rufulus) disturbance regimes under warming climate from 2000 to 2150 in the Central Hardwood Forests, USA. We quantified the effects size and relative importance of direct effects and vegetation feedback through a factorial design and two-way ANOVA. We found that vegetation feedback mediated the direct effects by altering primary host availability. Direct effects had more influence than vegetation feedback on ROB disturbed area in the short-term (e.g., 50 years), whereas vegetation feedback was more important in the mid- and long-term (e.g., 100-150 years). This was because the primary hosts of ROB, including northern red oak (Quercus rubra L.), southern red oak (Q. falcate Michx.), and black oak (Q. velutina Lam.), substantially declined over time during the simulation. Our study reveals how vegetation feedback mediates the warming effects on insect disturbance regimes and quantifies its impacts. We conclude that vegetation feedback is pivotal in understanding insect disturbance regimes under a warming climate as it may mediate or even reverse the expectation of increased insect disturbance. Without considering vegetation feedback, long-term predictions of insect disturbance made in ecological modeling studies may cause overestimation.