Supporting Information

Fig. S1 Flow chart of the article identification and selection process.

Fig. S2 Frequency distribution of invasive species’ trait values and CWM trait dissimilarities. Studies with an N-fixing invasive species are shown in black fill bars, whereas gray fill bars represent studies with non-fixing invasive species.

Fig. S3 Effect size funnel plots. A symmetrical funnel shape relationship between study effect sizes and standard errors suggests that there is no publication bias toward either large positive or negative effect sizes.

Fig. S4 Invasive species’ with higher leaf %N promote greater increases in soil N. Point size is scaled by variance in the study’s effect size such that larger points have smaller variance. Point shade represents CWM quality rank such that darker points are higher in quality. Solid line and blue shaded region show the meta-regression model fit and the 95% confidence interval. Dashed y=0 line aids in distinguishing between studies with positive (higher values in invaded area) and negative (lower values in invaded area) effect sizes.

Fig. S5 Study design mediates soil nitrate and organic matter invasion effect sizes. Different letters indicate significant differences among study designs according to post-hoc t-tests (alpha=0.05). Effect sizes were calculated as the standard mean difference (Hedges’ d) of invaded minus reference areas such that values >0 indicate greater values in invaded than reference areas; mean +/- 95% confidence interval with the number of studies shown in parentheses.

Table S1 List of the 143 articles used in our meta-analysis on the impacts of invasion on soil N cycling. If the article appears in a previous meta-analysis on this topic, there is a ‘Yes’ in one of the three previous meta-analysis columns: Liao 2008, Vila 2011, and Castro-Diez 2014. For inclusion in this meta-analysis, at least one study within each article needed to have provided data for one of the nine soil N measurement types listed. The number of studies per soil N measurement type that were extracted from each article is shown below.

See attached file: Table_S1.xlsx

Table S2 Summary of trait data quality by community and trait type. CWM values that were reported within the study have the highest quality with a quality rank of four. CWM value estimates that were calculated with more than 25% of the species in the community having the following attributes are considered higher in quality: a) measured species’ cover data (as opposed to cover estimate based on author’s description), b) species-specific cover data (as opposed to assuming species are present in equal abundance, c) species-specific trait values that are either reported within the original study or extracted from TRY.

Comm.
type / CWM calc’d / More than 25% of species in the community have... / Qual. rank / Number of CWM values / CWM values (%) /
Measured cover / Sp.-specific cover / Reported traits / Sp.-specific TRY traits / Leaf %N / Litter %N / Leaf C:N / Litter C:N / Leaf %N / Litter %N / Leaf C:N / Litter C:N /
Invasive sp. / No / NA / NA / NA / NA / 4 / 48 / 8 / 10 / 7 / 13.2 / 16.0 / 4.8 / 14.9
Yes / y / y / y / NA / 3 / 14 / 6 / 11 / 6 / 3.8 / 12.0 / 5.3 / 12.8
y / y / y / 3 / 26 / 0 / 1 / 2 / 7.1 / 0.0 / 0.5 / 4.3
y / y / 2 / 12 / 0 / 6 / 0 / 3.3 / 0.0 / 2.9 / 0.0
y / y / 2 / 6 / 0 / 0 / 0 / 1.6 / 0.0 / 0.0 / 0.0
y / y / NA / 2 / 45 / 19 / 39 / 14 / 12.4 / 38.0 / 18.8 / 29.8
y / y / 2 / 144 / 0 / 60 / 13 / 39.6 / 0.0 / 28.8 / 27.7
y / 1 / 0 / 2 / 4 / 0 / 0.0 / 4.0 / 1.9 / 0.0
y / 1 / 48 / 10 / 58 / 4 / 13.2 / 20.0 / 27.9 / 8.5
y / NA / 1 / 2 / 3 / 0 / 1 / 0.5 / 6.0 / 0.0 / 2.1
y / 1 / 19 / 0 / 15 / 0 / 5.2 / 0.0 / 7.2 / 0.0
0 / 0 / 2 / 4 / 0 / 0.0 / 4.0 / 1.9 / 0.0
Invaded area / No / NA / NA / NA / NA / 4 / 26 / 24 / 14 / 20 / 7.6 / 36.4 / 6.6 / 33.3
Yes / y / y / y / NA / 3 / 16 / 6 / 12 / 7 / 4.7 / 9.1 / 5.7 / 11.7
y / y / y / 3 / 26 / 0 / 1 / 0 / 7.6 / 0.0 / 0.5 / 0.0
y / y / 2 / 7 / 0 / 6 / 0 / 2.0 / 0.0 / 2.8 / 0.0
y / y / NA / 2 / 4 / 0 / 1 / 0 / 1.2 / 0.0 / 0.5 / 0.0
y / y / 2 / 9 / 0 / 0 / 1 / 2.6 / 0.0 / 0.0 / 1.7
y / y / NA / 2 / 37 / 15 / 34 / 11 / 10.8 / 22.7 / 16.0 / 18.3
y / y / 2 / 133 / 0 / 55 / 13 / 38.9 / 0.0 / 25.9 / 21.7
y / 1 / 2 / 2 / 4 / 1 / 0.6 / 3.0 / 1.9 / 1.7
y / 1 / 30 / 10 / 55 / 4 / 8.8 / 15.2 / 25.9 / 6.7
y / NA / 1 / 6 / 7 / 3 / 2 / 1.8 / 10.6 / 1.4 / 3.3
y / 1 / 46 / 0 / 23 / 1 / 13.5 / 0.0 / 10.8 / 1.7
0 / 0 / 2 / 4 / 0 / 0.0 / 3.0 / 1.9 / 0.0
Reference area / No / NA / NA / NA / NA / 4 / 25 / 24 / 14 / 20 / 7.8 / 40.7 / 9.0 / 37.0
Yes / y / y / y / NA / 3 / 16 / 5 / 12 / 6 / 5.0 / 8.5 / 7.7 / 11.1
y / y / y / 3 / 23 / 0 / 0 / 0 / 7.2 / 0.0 / 0.0 / 0.0
y / y / 2 / 5 / 0 / 5 / 0 / 1.6 / 0.0 / 3.2 / 0.0
y / y / NA / 2 / 4 / 1 / 0 / 1 / 1.3 / 1.7 / 0.0 / 1.9
y / y / 2 / 9 / 0 / 1 / 1 / 2.8 / 0.0 / 0.6 / 1.9
y / y / NA / 2 / 32 / 12 / 23 / 8 / 10.0 / 20.3 / 14.8 / 14.8
y / y / 2 / 94 / 0 / 9 / 0 / 29.4 / 0.0 / 5.8 / 0.0
y / 1 / 4 / 0 / 4 / 1 / 1.3 / 0.0 / 2.6 / 1.9
y / 1 / 25 / 5 / 47 / 6 / 7.8 / 8.5 / 30.3 / 11.1
y / NA / 1 / 15 / 8 / 14 / 5 / 4.7 / 13.6 / 9.0 / 9.3
y / 1 / 59 / 1 / 21 / 2 / 18.4 / 1.7 / 13.5 / 3.7
0 / 9 / 3 / 5 / 4 / 2.8 / 5.1 / 3.2 / 7.4

Table S3 Impact of invasion on soil N cycling with associated publication bias checks and Q statistics. See Methods for a description of the fail-safe number, correlation among effect and sample size, and Q statistics.

Soil measurement / Grand effect size / Fail-safe number / Correlation: Effect and sample size / Q statistics
Effect size (d+) / 95% CI / k / num. / p-value / r / slope coef. / p-value / QE / QE p-value / QM / QM p-value
Inorganic N pool / NH4+ / 0.26 / 0.12 to 0.40 / 141 / 1827 / <.001 / 0.01 / -0.01 / ns / 307.9 / <.001 / 13.4 / <.001
NO3- / 0.17 / -0.003 to 0.34 / 150 / 314 / <.01 / -0.17 / -0.01 / ns / 405.2 / <.001 / 3.7 / 0.054
NH4++NO3- / 0.15 / -0.05 to 0.35 / 195 / 404 / <.01 / -0.06 / -0.01 / ns / 492.2 / <.001 / 2.0 / 0.153
N flux / Δ NH4+ / -0.05 / -0.31 to 0.20 / 50 / 0 / ns / 0.07 / 0.01 / ns / 74.5 / 0.011 / 0.2 / 0.699
Δ NO3- / 0.45 / 0.21 to 0.69 / 79 / 1321 / <.001 / -0.33 / -0.03 / ns / 167.0 / <.001 / 13.4 / <.001
Δ NH4++NO3- / 0.36 / 0.14 to 0.58 / 101 / 1143 / <.001 / -0.19 / -0.07 / 0.046 / 207.9 / <.001 / 10.0 / 0.002
Other / Soil N / 0.53 / 0.31 to 0.75 / 187 / 8416 / <.001 / -0.01 / -0.01 / ns / 608.3 / <.001 / 21.7 / <.001
Soil C:N / -0.05 / -0.25 to 0.14 / 100 / 0 / ns / 0.06 / 0.01 / ns / 190.7 / <.001 / 0.3 / 0.586
SOM / 0.59 / 0.18 to 0.99 / 90 / 1508 / <.001 / 0.05 / -0.02 / ns / 368.5 / <.001 / 8.0 / 0.005

Table S4 Mixed effects model results that included trait dissimilarity as a significant fixed effect (Invaded – Reference CWM) and the results of re-running that model with trait dissimilarity calculated as the invasive species trait value minus the reference CWM (Invasive sp. – Reference CWM). Random effects account for variation within studies and among studies within the same article. For each model fit, the following data are reported: fixed-effect coefficient (est), pseudo-R2 (Rp2), and number of studies (k). Models with fixed effect p-values less than alpha=0.1 are highlighted in grey.

Response Var / Trait type / Invaded - Reference CWM / Invasive sp. - Reference CWM
est / Rp2 / k / est / Rp2 / k
Δ NH4++NO3- / Leaf C : N / -0.03 / 0.00 / 46 / -0.04 / 0.00 / 43
Soil C:N / 0.03 / 0.00 / 66 / 0.02 / 0.00 / 62
NH4++NO3- / Litter %N / 1.55 / 0.72 / 24 / 0.53 / 0.28 / 17
NH4++NO3- / Litter C : N / -0.02 / 0.35 / 23 / -0.01 / 0.00 / 14

Methods S1 Hedges’ d calculation.

Hedge’s d for a given study was calculated as d=(Xinv-Xref )S J, where Xinv and Xref are soil measurement means from invaded and reference areas, respectively, for each study. S is the pooled standard deviation and J is a weighting factor based on the number of replicates such that J=1- 34Ninv+Nref-2-1 . The variance for Hedges’ d was calculated as vard=Ninv+NrefNinvNref+ d22Ninv+Nref .

Methods S2 Pseudo-R2 calculation.

A pseudo-R2 was calculated as the proportional reduction in the total variance such that (σred., article2+ σred., study2)+(σfull, article2+ σfull, study2)(σred., article2+ σred., study2) , where σred., article2 and σred., study2 are the variances associated with the nested random effects, article and study, from a reduced model that does include fixed effects. The terms σfull, article2 and σfull, study2 are the variances associated with the nested random effects from the full model.

Notes S1 Invasion impacts on soil nitrogen cycling references.

Abella SR, Chiquoine LP, Backer DM. 2012. Ecological Characteristics of Sites Invaded by Buffelgrass (Pennisetum ciliare). Invasive Plant Science and Management 5: 443–453.

Adair CE, Burke IC. 2010. Plant phenology and life span influence soil pool dynamics: Bromus tectorum invasion of perennial C3–C4 grass communities. Plant and Soil 335: 255–269.

Adams SN, Engelhardt KAM. 2009. Diversity declines in Microstegium vimineum (Japanese stiltgrass) patches. Biological Conservation 142: 1003–1010.

Aguilera AG, Alpert P, Dukes JS, Harrington R. 2010. Impacts of the invasive plant Fallopia japonica (Houtt.) on plant communities and ecosystem processes. Biological Invasions 12: 1243–1252.

Allison SD, Nielsen C, Hughes RF. 2006. Elevated enzyme activities in soils under the invasive nitrogen-fixing tree Falcataria moluccana. Soil Biology and Biochemistry 38: 1537–1544.

Archibald RD, Craig MD, Bialkowski K, Howe C, Burgess TI, Hardy GESJ. 2011. Managing small remnants of native forest to increase biodiversity within plantation landscapes in the south west of Western Australia. Forest Ecology and Management 261: 1254–1264.

Ashton IW, Hyatt LA, Howe KM, Gurevitch J. 2005. Invasive species accelerate decomposition and litter nitrogen loss in a mixed deciduous forest. Ecological Applications 15(4): 1263-1272

Asner GP, Beatty SW. 1996. Effects of an African grass invasion on Hawaiian shrubland nitrogen biogeochemistry. Plant and Soil 186: 205–211.

Baer SG, Church JM, Williard KWJ, Groninger JW. 2006. Changes in intrasystem N cycling from N2-fixing shrub encroachment in grassland: multiple positive feedbacks. Agriculture, Ecosystems & Environment 115: 174–182.

Bansal S, Sheley RL, Blank B, Vasquez EA. 2014. Plant litter effects on soil nutrient availability and vegetation dynamics: changes that occur when annual grasses invade shrub-steppe communities. Plant Ecology 215: 367–378.

Belnap J, Phillips SL. 2001. Soil biota in an ungrazed grassland: response to annual grass (Bromus tectorum) invasion. Ecological Applications 11: 1261–1275.

Belnap J, Phillips SL, Sherrod SK, Moldenke A. 2005. Soil biota can change after exotic plant invasion: does this affect ecosystem processes? Ecology 86: 3007-3017.

Blank RR. 2008. Biogeochemistry of Plant Invasion: A Case Study with Downy Brome (Bromus tectorum). Invasive Plant Science and Management 1: 226–239.

Blank RR. 2010. Intraspecific and interspecific pair-wise seedling competition between exotic annual grasses and native perennials: plant-soil relationships. Plant and Soil 326: 331–343.

Blank RR, Young JA. 2002. Influence of the exotic invasive crucifer, Lepidium latifolium, on soil properties and elemental cycling. Soil Science 167: 821-829.

Booth MS, Stark JM, Caldwell MM. 2003. Inorganic N turnover and availability in annual-and perennial-dominated soils in a northern Utah shrub-steppe ecosystem. Biogeochemistry 66: 311-330

Boswell CC, Espie PR. 1998. Uptake of moisture and nutrients by Hieracium pilosella and effects on soil in a dry sub‐humid grassland. New Zealand Journal of Agricultural Research 41: 251-261.

Bottollier-Curtet M, Charcosset J-Y, Poly F, Planty-Tabacchi A-M, Tabacchi E. 2012. Light interception principally drives the understory response to boxelder invasion in riparian forests. Biological Invasions 14: 1445–1458.

Boudiaf I, Baudoin E, Sanguin H, Beddiar A, Thioulouse J, Galiana A, Prin Y, Le Roux C, Lebrun M, Duponnois R. 2013. The exotic legume tree species, Acacia mearnsii, alters microbial soil functionalities and the early development of a native tree species, Quercus suber, in North Africa. Soil Biology and Biochemistry 65: 172–179.