Influence of multiple global change drivers on plant invasion: Additive effects are uncommon

Abstract

Invasive plants threaten biodiversity and cause huge economic losses. It is thought that global change factors (GCFs) associated with climate change (including shifts in temperature, precipitation, nitrogen, and atmospheric CO₂) will amplify their impacts. However, only few studies assessed mixed factors on plant invasion. We collated the literature on plant responses to GCFs to explore independent, combined, and interactive effects on performance and competitiveness of native and invasive plants. From 176 plant species, our results showed that: (1) when native and invasive plants are affected by both independent and multiple GCFs, there is an overall positive effect on plant performance, but a negative effect on plant competitiveness; (2) under increased precipitation or in combination with temperature, most invasive plants gain advantages over natives; and (3) interactions between GCFs on plant performance and competitiveness were mostly synergistic or antagonistic. Our results indicate that native and invasive plants may be affected by independent or combined GCFs, and invasive plants likely gain advantages over native plants. The interactive effects of factors on plants were non-additive, but the advantages of invasive plants may not increase indefinitely. Our findings show that inferring the impacts of climate change on plant invasion from factors individually could be misleading. More mixed factor studies are needed to predict plant invasions under global change.

Keywords: climate change; effect sizes; interactions; nitrogen deposition; precipitation; temperature.

Figure 1

An illustration of the system for classifying plant responses to multiple global change factors based on the direction and magnitude of their individual and combined effects. According to the interaction type of two factors with double positive, double negative and opposite individual effects (see Figure 1 of Crain et al. 2008), the interaction type (A) of “limiting synergy” proposed by Piggott et al. (2015a) under opposite individual effects. In this study, interaction classification criteria for double positive (B), double negative (C) and opposite (D) individual effects. In the figure, bold numbers denote net trait response from factor A, B or A+B, numbers in brackets denote antagonistic or synergistic effects.

Figure 2

Schematic of the bootstrap resampling design for estimating the joint effect sizes of single and multiple GCFs treatments on performance and competitiveness of native and invasive plants using plant performance as an example. In step 1, according to no treatment (P_C , sample size of data is N_C ), A factor alone (P_A , sample size of data is N_A ), B factor alone (P_B , sample size of data is N_B ) and AB two-factor interaction (P_AB , sample size of data is N_AB ) standardized plant performance data. From steps 2 to 6, it is repeated 1000 times, and the number of iterations is marked with i in subscript form. In step 2, 100 samples from P_C , P_A , P_B and P_AB were randomly placed back to draw each as P_Ci , P_Ai , P_Bi and P_ABi , and in step 3, the mean values of 100 random samples were calculated as M_Ci , M_Ai , M_Bi and M_ABi , the single A factor (E _Ai ), the single B factor (E_Bi ) and the A, B two-factor combination (E_ABi ) are calculated in step 4 by calculating the difference between the average value under the obtained treatment and the average value of the control group. Effect size for i iterations. Step 5, estimate the predicted two-factor additive effect size (AD_ABi ) based on the effect size of the single factors A and B, and compare the difference between the actual effect size of the two factors A and B and the predicted additive value (D_ABi ) in step 6. Step 7, calculate the mean and 2.5% and 97.5% quantiles of the effect size and difference size obtained by 1000 iterations, as the lower and upper interval of the combined effect (difference) size and confidence interval. In step 8, interactions are identified according to the interaction classification criteria and plotted against the data in a final step 9.

Figure 3

Independent effects of single GCFs on plant invasion in: (A) plant performance (left panel) and (B) competitiveness (right panel). Values indicate the means with 95% confidence intervals (CIs), and original sample size numbers for native and invasive plants are shown on the right hand side of dots. The * on the left of the means indicates the effect size is significantly different from zero (CI does not overlap with 0). A significant positive effect is where the CI is greater than 0 and a negative effect is where the CI is less than 0. Factors represented in the ordination: T = temperature increase, P = precipitation increase, D = drought intensification, N = nitrogen deposition, C = atmospheric carbon dioxide enrichment, GCFs (S) = single global change factors summary

Figure 4

Combined effects of multiple GCFs on plant invasion in (A) plant performance (left panel) and (B) competitiveness (right panel). Values indicate the means with 95% confidence intervals (CIs), and original sample size numbers for native and invasive plants are shown right the dots. The * to the left of the dots indicates that CI does not overlap zero, which means that the GCF has a significant impact on plants, and positive effect is where the CI is greater than zero, or a negative effect where the CI is lower than zero. TD: temperature increase × drought intensification, TN: temperature increase × nitrogen deposition, TC: temperature increase × atmospheric carbon dioxide enrichment, PN: precipitation increase × nitrogen deposition, DN: drought intensification × nitrogen deposition, DC: drought intensification × atmospheric carbon dioxide enrichment, NC: nitrogen deposition × atmospheric carbon dioxide enrichment, GCFs(M): multiple global change factors summary.

Figure 5

Interactive effects of multiple GCFs on invasive plants: (A) plant performance and (B) competitiveness. Values indicate the means with 95% confidence intervals (CIs) in left panel. The * to the left of the dots indicates that CI does not overlap with zero meaning that the GCF has a significant interaction effect on plants. Values in percentages in indicate the proportions of the type of interaction effect among factors. Additive effects are cases where the effect under multiple GCFs is the same as the sum of those effects individually. Antagonistic and synergistic effects (unsigned) are cases where both factors have effects in the same direction, while signed effects (positive and negative) are weaker or stronger than expected. Details on classifications of these interactive effects are in the text. The abbreviations for the factors are the same as those in  Figure 5 .