Spatial characteristics of an Amur honeysuckle (Lonicera maackii (Rupr.) Herder) invasion in an Indiana mixed-hardwoods forest

Background/Question/Methods

Amur honeysuckle (Lonicera maackii (Rupr.) Herder) is a non-native, invasive shrub that has aggressively colonized many forests in the central U.S.  While it is well known that this species suppresses the growth and regeneration of native plants, less is known about factors influencing the spatial distribution and life stage characteristics of an invading population.  During the fall of 2011, we recorded the locations of every Amur honeysuckle individual ≥0.5 m tall and every tree ≥10 cm diameter at breast height (dbh) in a 60 m x 60 m area within a mixed-hardwoods forest in West Lafayette, Indiana.  For each Amur honeysuckle, we also recorded diameter above root burl (cm), height (m), and whether it produced berries.  Ripley’s K function, χ² quadrat test, and the Clark-Evans nearest-neighbor method were used to test whether the spatial distribution of Amur honeysuckle individuals was different from random.  Ripley’s cross-Kwas used to test whether spatial independence existed between immature Amur honeysuckles (those without berries) and mature honeysuckles (those with berries) and between Amur honeysuckles and overstory trees.  Regression analysis incorporating spatial correlation was used to examine factors influencing the diameter and height of both immature and mature Amur honeysuckle individuals.

Results/Conclusions

We recorded the spatial locations of 2,386 Amur honeysuckle individuals and 231 overstory trees.  The quadrat test indicated that Amur honeysuckle individuals were not randomly distributed in two-dimensional space (χ² = 913.6, p < 0.001).  Both the Ripley’s K function and the Clark-Evans method indicated that Amur honeysuckle individuals exhibited a clustered spatial pattern (estimated K(r) > theoretical K(r); Clark-Evans R = 0.88).  Results from Ripley’s cross-K indicated that immature individuals tended to be located near mature individuals and Amur honeysuckle plants tended to be located near overstory trees (i.e., no spatial independence or repulsion between Amur honeysuckle plants of different reproductive status or between Amur honeysuckle and trees).  All regression analyses that examined factors influencing Amur honeysuckle diameter and height were significant (p < 0.05).  Incorporating spatial correlation in the regression models yielded a better fit (based on Akaike information criterion [AIC]) than when spatial correlation was not considered.  In some cases, predictor variables that were otherwise significant became non-significant (p > 0.05) when spatial correlation was considered in the model.  Future work will incorporate additional study sites with lower densities of Amur honeysuckle than the site examined here.