Garlic mustard’s (Alliaria petiolata) effectiveness as an invader of Eastern North American deciduous forest groundlayers

Background/Question/Methods

Several studies indicate garlic mustard may be more of an opportunist than a problematic invasive species and its abundance and effectiveness as an invader has recently declined.  To determine garlic mustard impact on native species, we established 240 2.5 x 2.5 m treatment plots in 2004 from which second-year garlic mustard plants were removed annually beginning in 2005 and continuing to the present.   Plots were randomly assigned to control, early (mid-March) or late (early-May) removal of second-year garlic mustard.  Data were collected on first- and second-year garlic mustard density and aerial cover was estimated for all vascular plants in a 50 x 50 cm sampling plot centrally located in each treatment plot.  We used repeated measure MANOVA to examine change in cover of first-year garlic mustard plants in treatments across the study period.  For 2004 (pretreatment) and 2011 samples, nonmetric multidimensional scaling (NMS) was used to ordinate plots using cover of native species.  NMS axes scores and native species total cover were used in MAVOVA to determine treatment effects.  We tested for differences in the portion of sample plots without garlic mustard among treatments using chi-square.

Results/Conclusions

First-year garlic mustard percent cover did not differ significantly (p<0.603) among treatments in 2004 or 2005 and was highest for all treatments in 2005 being (mean±SE) 33.5±3.0, 35.8±3.4, and 28.7±2.7 for control, early, and late treatments, respectively.  After 2006, there were no significant (p<0.05) differences between early and late removal treatments, but both had significantly (p<0.001) lower first-year plant cover than the control.  First-year plant cover declined precipitously after 2007 and in 2011 reached lows of 1.1±0.2, 0.2±0.1, and 0.1±0.04 for control, early, and late treatments, respectively.  Proportion of plots without garlic mustard in 2004 was low (0.012 to 0.062) and did not differ significantly (p<0.25) among treatments.  However, in 2011 early (0.67) and late (0.48) treatments had a significantly (p<0.001) higher proportion of plots without garlic mustard than the control (0.09).  For 2004, there were no significant (p<0.5450) treatment effects on native species total cover, and NMS axes scores did not differ in 2004 or 2011.  However, in 2011 early (66.8±1.8) and late (65.3±2.2) treatments had significantly (p<0.001) higher cover than the control (55.1±1.9). Our results suggest garlic mustard’s decline in abundance occurred over a short time and even at low abundances this invader’s legacy persists.