Effects of Defoliation and Herbivore Exclosures on Growth and Reproduction of Transplanted Bunchgrass Seedlings

Effects of Defoliation and Herbivore Exclosures on Growth and Reproduction of Transplanted Bunchgrass Seedlings

Justin M. Valliere (La Kretz Center for California Conservation Science, Institute of the Environment and Sustainability, University of California, Los Angeles, Los Angeles, CA 90095, justin.valliere{at}uwa.edu.au)

Grasslands in California have been heavily impacted by human activities, threatening their high biodiversity and important economic value (Murphy and Ehrlich 1989). As such, these ecosystems are targets for restoration, with the establishment of native perennial bunchgrasses often being the primary focus (Stromberg et al. 2007). One method of bunchgrass reestablishment is the transplanting of nursery-propagated seedlings. However, the performance and survival of transplanted individuals may be negatively impacted by several factors, including herbivory. Methods that improve plant growth and herbivory tolerance and resistance could therefore aid in the restoration of this plant community.

Multiple studies have explored the use of defoliation methods such as grazing, clipping, and mowing for the restoration of California grasslands. These studies yielded differing results. For example, Valliere et al. (2019) found that mowing increased cover of the native bunchgrass Stipa pulchra and the number of seeds of this species in the seed bank. In contrast, Hayes and Holl (2003) and Lulow (2008) found no benefit of grazing or clipping, respectively, on native bunchgrasses, while Kimball and Schiffman (2003) reported negative effects of clipping. Because these studies applied treatments at the community-level, it is not possible to discern direct effects of defoliation on individual plants from indirect effects via plant neighbors.

Most grass species display some level of herbivory tolerance, and while the loss of foliage may initially reduce plant growth, there may be longer-term benefits that could improve restoration success. Some grasses may respond with compensatory growth and increased photosynthesis following defoliation (Nowak and Caldwell 1984), and some exhibit greater growth or “overcompensation” compared to ungrazed plants (Paige and Whitham 1987). Herbivory may also induce plant responses that reduce the performance and preference of herbivores, ultimately enhancing plant fitness (Agrawal 1998). Simulated herbivory (e.g., clipping) may have similar beneficial effects, and therefore clipping is one method by which practitioners can increase the hardiness of plants used for restoration (Amme 1985).

In two separate experiments, I evaluated the effects of clipping and herbivore exclosures on the growth of native bunchgrass seedlings used in a grassland restoration project in southern California. I explored two hypotheses: (1) simulated herbivory (i.e., clipping) would result in increased growth of transplanted seedlings due to compensatory responses; and (2) simulated herbivory prior to transplanting would increase plant tolerance and resistance to herbivory later in the field.

In Experiment 1, I tested the effects of defoliation on the compensatory growth of ten native bunchgrass species: Agrostis pallens (leafy bentgrass), Elymus condensatus (giant wild rye), Koeleria macrantha (prairie junegrass), Melica californica (California melicgrass), Melica imperfecta (small-flowered melicgrass), Muhlenbergia rigens (deergrass), Stipa cernua (nodding needlegrass), Stipa coronata (crested needlegrass), Stipa lepida (foothill needlegrass), and Stipa pulchra (purple needlegrass). Plants were established from seed in 160 mL conical pots (one plant per pot) in November 2017 at the University of California, Los Angeles. Plants were watered 1–2 times per week. I grew plants for three months, at which point they were transplanted to the field at the University of California’s Stunt Ranch Reserve, located in southern California’s Santa Monica Mountains. The restoration site was a disturbed grassland consisting of mostly exotic annuals. Prior to transplanting, half of the replicates for each species (n = 10) were clipped at ~5 cm above soil-level. Cylindrical exclosures (~15 cm in diameter) of 1 cm² hardware cloth were staked around all individuals. These cages protected plants against larger-sized herbivores (e.g., mammals) but not insect herbivory. I watered plants following transplanting, then once every two weeks for the following six weeks, and then once per month until the end of May 2018. In Experiment 2, I tested the effects of both defoliation and the presence of herbivore exclosures on the growth of transplanted S. pulchra seedlings. I chose this species because it is often the most abundant bunchgrass in California grasslands (Bartolome and Gemmill 1981). Plants were grown under the same conditions as described above. In a full factorial design, replicates (n = 15) were either clipped or left intact prior to transplanting and were either enclosed in hardware cloth or left exposed.

For each experiment, I counted plant inflorescences and collected shoot biomass in June 2018. Biomass was dried at 60°C for 48 hours prior to weighing. For Experiment 1, I used analysis of variance (ANOVA) to analyze biomass data, with species, treatment (clipped or unclipped), and the interaction included as fixed effects. I used Student’s t-tests to compare the biomass of clipped and unclipped plants within species. To analyze inflorescence data, I used a generalized linear model (GLM) with a Poisson distribution, with species, treatment, and the interaction included as fixed effects and individual chi-squared tests by species. For Experiment 2, I again used ANOVA and GLM to analyze biomass and inflorescence data, respectively, with clipping and exclosure treatments and the interaction as fixed effects. I then used Tukey’s HSD test for post hoc mean comparisons. All analyses were performed in R (RStudio v. 1.1.419, R v. 3.5.2, R Core Team 2018).

For both experiments, survival was high (> 93%), and most plant mortality was due to gopher activity. In Experiment 1, biomass (Figure 1A) varied by species, clipping treatment, and the interaction of species and treatment. Four species showed a significant growth response to defoliation; A. pallens, S. cernua, and S. lepida showed a negative response, while E. condensatus, showed a positive response (Table 1, Figure 1A). Inflorescence number (Figure 1B) varied by species and treatment but not the interaction. Two species, K. macrantha and S. pulchra showed a positive effect of clipping on inflorescence number (Table 1, Figure 1B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Mean (± SE) shoot mass (A) and inflorescence number (B) from Experiment 1 testing the effects of clipping (control = light bars, clipped = dark bars) on transplanted seedlings of ten perennial grass species (n = 10). Asterisks indicate a significant difference between previously clipped and un-clipped plants (control) within each species. Presence of asterisks indicates statistical significance: *p < 0.05; **p < 0.01; ***p < 0.0001.

In Experiment 2, clipping treatment and the presence of exclosures, but not the interaction, had a significant effect on biomass of S. pulchra (Figure 2A). Plants that were unclipped and exposed (i.e., no herbivore exclosures) had significantly lower biomass than plants in all other treatments. There was no difference in biomass between clipped and unclipped plants that were protected by exclosures. Clipping led to a greater number of inflorescences produced by S. pulchra plants, but the presence of herbivore exclosures or the interaction between the two treatments had no effect on reproductive output (Figure 2B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Mean (± SE) shoot mass (A) and inflorescence number (B) from Experiment 2 testing the effects of clipping (control = light bars, clipped = dark bars) and the presence of herbivore exclosures (exposed or caged) on transplanted Stipa pulchra seedlings in a full-factorial experiment (n = 15). Different letters above bars show significant differences by treatment. Presence of asterisks indicates statistical significance: *p < 0.05; **p < 0.01; ***p < 0.0001.

The results of Experiment 1 show that plant responses to defoliation prior to transplanting are species-specific. In the first experiment, biomass of most species showed either a negative or neutral response to the clipping treatment, with only a single species, E. condensatus, exhibiting a positive response in terms of shoot biomass. These responses could indicate differing levels of tolerance to the simulated herbivory treatment, as species differ in their ability to store and reallocate resources, grow new tissue, and increase photosynthetic efficiency of remaining foliage (Hamerlynck et al. 2016). Additionally, the different species-level responses I observed could also be related to differing phenologies among species. For example, E. condensatus flowers later than many of the other cool-season grasses included in this study (Vorobik 2012) and may remain active during the summer dry period when other species senesce. The ability of this species to continue growth into summer months could have contributed to the positive response observed.

Even in the few species that showed reduced biomass due to earlier clipping, the differences between treated and untreated plants were rather minimal, highlighting the ability of these bunchgrass species to recover from even extreme defoliation. Additionally, despite the negative growth response observed in some species, none showed negative effects on inflorescence development. Two species even showed positive effects of clipping on reproductive output.

Perhaps the most interesting result is the benefit of clipping on growth of S. pulchra in Experiment 2 only manifested in the absence of herbivore exclosures. This species showed no significant differences in shoot biomass between clipped and unclipped plants that were protected by wire cages, but clipped plants that were not protected from herbivores in Experiment 2 grew much larger than unclipped controls. Unclipped plants may have been more attractive to herbivores, possibly because of induced responses in treated plants (Agrawal 1998). Clipped plants may have also been more tolerant to further defoliation after being transplanted. There is substantial evidence that exposure to stress events such as herbivory can improve plant responses to future stressors (Walter et al. 2013). In some cases, practitioners may be able to exploit this “ecological stress memory” (Walter et al. 2013) to improve restoration success (Valliere et al. 2019).

Is clipping prior to transplanting a useful method for the restoration of these bunchgrass species? These results suggest that some species may benefit from such a treatment, but these responses are highly species-specific, and some species may also be negatively affected. Practitioners should conduct pilot studies to identify how sensitive target species are to defoliation prior to transplanting and determine if any may benefit. It is possible that more species would have shown positive responses if exposed to herbivores (i.e., not protected by exclosures), as S. pulchra did in the second experiment. Future work should explore this potential stress-conditioning phenomenon. The effects of this treatment will likely depend on the timing and severity of the treatment. Seedlings may also be more sensitive to defoliation than mature individuals. Longer-term monitoring and similar experiments with adult plants will be useful in further evaluating the potential benefits (or drawbacks) of this method for grassland restoration.