Functional form and interactions of the drivers of understory non‐native plant invasions in northern US forests. Issue 12 (3rd October 2019)

• Record Type: Journal Article
• Authors: Golivets, Marina; Woodall, Christopher W.; Wallin, Kimberly F.
• Editors: Pauchard, Anibal
• Is Part Of: Journal of applied ecology. Volume 56:Issue 12(2019)
• Page Start: 2596
• Page End: 2608
• Publication Date: 2019-10-03
• Subjects: Bayesian -- biological invasions -- boosted regression trees -- drivers of invasion -- forest understory -- non‐native plants -- spatial scale; Agriculture -- Periodicals; Biology, Economic -- Periodicals; Agricultural ecology -- Periodicals; Applied ecology -- Periodicals; 577
• DOI: 10.1111/1365-2664.13504
• Languages: English
• ISSNs: 0021-8901

Abstract

Abstract: The number and rate of non‐native plant invasions in forests have been steadily increasing over the last century with profound consequences for the composition, structure and functioning of these ecosystems. While multiple regional, landscape and local environmental factors are known to drive the spread of non‐native invasive plant species (NNIPS) into forests, such factors have rarely been analysed within a unified analytical framework allowing for the assessment of their relative importance, possible non‐linear behaviour and interactions. Obtaining such knowledge would improve the understanding and allow prediction of forest plant invasions and help prioritize actions for effective NNIPS control and mitigation. We used boosted regression trees and Bayesian non‐linear regression to analyse forest inventory data spanning 14 northern US states in combination with data on climate, land use and disturbance. Within each framework, we assessed the magnitude, direction and functional form of the effects of 18 regional to local environmental factors and their interactions on the probability of presence and richness of NNIPS in the forest understory, while controlling for sampling intensity, data collection time and spatial autocorrelation. The studied drivers together captured nearly 50% of the variation in NNIPS occurrence and richness and were sufficient to obtain sensible predictions. Most drivers exhibited monotonic non‐linear individual effects and several drivers showed non‐additive combined effects. Temperature and landscape openness were the two major determinants of NNIPS presence and richness, but their effects were to a different degree moderated by elevation, climatic and topographic wetness, and stand age. Stand age was a more important predictor of NNIPS invasion than stand productivity. Deer density tended to be positively associated with NNIPS, whereas gypsy moth outbreaks possibly have increased resistance to plant invasions. Synthesis and applications. To optimize non‐native invasive plant species (NNIPS) control efforts in forests, it is important to explicitly account for the functional form and interactions among the factors operating across spatial and temporal scales. Increasing the average stand age at the landscape and regional levels is essential for buffering against further invasions. Recurrent canopy disturbance likely slowed down NNIPS invasion previously, yet whether this holds under higher propagule pressure remains unknown. Abstract : To optimize non‐native invasive plant species (NNIPS) control efforts in forests, it is important to explicitly account for the functional form and interactions among the factors operating across spatial and temporal scales. Increasing the average stand age at the landscape and regional levels is essential for buffering against further invasions. Recurrent canopy disturbance likely slowed down NNIPS invasion previously, yet whether this holds under higher propagule pressure remains unknown.