Ecological Archives A024-028-A4

Basil V. Iannone III, Moira L. Zellner, David H. Wise. 2014. Modeling the impacts of life-history traits, canopy gaps, and establishment location on woodland shrub invasions. Ecological Applications 24:467–483. http://dx.doi.org/10.1890/13-0833.1

Appendix D. Results pertaining to interactions among levels of age of first reproduction, fecundity, dispersal, and initial establishment location.

INTERACTIONS AMONG LEVELS OF AGE OF FIRST REPRODUCTION AND FECUNDITY

Age at first reproduction and fecundity interacted in how they affected all response variables except lag (Table 3 of manuscript), although these interactions were weak. The decrease in time50caused by increasing fecundity from low to high only varied from 42% to 47% across late and early age at first reproduction (Fig. D1A). The decrease in propreprocaused by the same increase in fecundity only varied from 0.19 to 0.15 across late and early age at first reproduction (Fig. D1B). The decrease in area:edge ratio caused by the same increase in fecundity only varied from 19% to 13% across late and early age at first reproduction (Fig. D1C).

INTERACTIONS INVOLVING GAP-DEPENDENT AGE AT FIRST REPRODUCTION AND FECUNDITY

Interpreting statistically significant interactions involving gap-dependence is not straightforward because gap-dependence is only one of three levels for each life-history trait. A substantial component of the interaction involves the differing degree to which the values for response variables generated by gap dependence are intermediate between those for the extreme values under constant environmental conditions. Inspection of Fig. D1 reveals that these interactions are relatively small compared to the difference between values yielded by the two extremes of each life-history trait under constant conditions.

INTERACTION INVOLVING DISPERSAL

Although rates of spread (time50) did not differ between invasions with random and gap-dependent dispersal, our analysis detected an extremely small, but statistically significant, interaction between levels of age at first reproduction and dispersal (Table 3A of manuscript). Gap-dependent dispersal caused time50 to decrease by 2 years when age at first reproduction was late and increase by 2.6 years when it was gap-dependent. Follow-up simulations comparing random and gap-dependent dispersal across the range of canopy-gap formation found in forests (from 1% to 2% per year) (Runkle 1982), while assuming gap-dependent age at first reproduction and fecundity (2 × 2 factorial design; N = 40), revealed that the small effect sizes of gap-dependent dispersal relative to random dispersal on invasion patterns prevailed regardless of rates of canopy-gap formation. Gap-dependent dispersal decreased propreprofrom 0.66 ± 0.005 to 0.64 ± 0.004 when gaps formed at a rate of 1% per year, but increased proprepro from 0.63 ± 0.006 to 0.64 ± 0.004 when gaps formed at a rate of 2% per year (ANOVA analysis; F1, 36 = 7.800; P = 0.008). Gap-dependent dispersal also caused invasions to become 6% more scattered (area:edge ratio decreased from 3.2 ± 0.07 to 3.0 ± 0.05), but only when gaps formed at a rate of 2% per year; although this interaction was only marginally significant (ANOVA results; F1, 36 = 3.412; P = 0.073). Varying rates of canopy-gap formation had no effect on how gap-dependent dispersal affected time50 or lag (ANOVA results; F1, 36 = 2.126 and 1.320; P = 0.15 and 0.26, respectively).

INTERACTIONS INVOLVING INITIAL ESTABLISHMENT LOCATION

The interaction between initial establishment location and life-history traits was statistically significant for three response variables (Table 3A – 3C of manuscript). The effect sizes of these interactions were small compared to the magnitude of the main effects (Fig. D2).

FigD1

Fig. D1. Interactions between age at first reproduction and fecundity in how variation in these life-history traits affected (A) rates of spread (time50), (B) proportion of the invasion reproducing (proprepro), and (C) spatial arrangement (area:edge ratio). N = 40 runs for each combination of age at first reproduction and fecundity (i.e., each bar). Numbers above bars in A and C signify the percentage decrease in time50 and area:edge ratio for that treatment relative to the treatment within the same level of age at first reproduction that had the greatest value. In B, because the response variable is already expressed as a proportion, the number above the bar is not percentage change, but is the amount of decrease in that treatment relative to the treatment within the same level of age of first reproduction with the highest value. Mean ± SE.


FigD2

Fig. D2. Effects of initial establishment location on rates of spread (time50) within levels of (A) age at first reproduction and (B) fecundity; and the effects of initial establishment location on (C) proportion of the invasion reproducing (proprepro) within levels of fecundity. N = 60 for each treatment level (bar). Percentages above bars in A and B signify the amount that time50 decreased in that treatment relative to the treatment within the same level of age at first reproduction or fecundity with the higher value. Proportions above bars in C signify the amount that proprepro decreased in that treatment relative to the treatment within the same level of fecundity with the highest value. Mean ± SE.


[Back to A024-028]