University of Nevada, Reno Succession in a post-fire world: Bunchgrass seedling dynamics after wildfire in sagebrush steppe ecosystems A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Natural Resources and Environmental Science by Jeffrey Gicklhorn Dr. Beth Newingham / Thesis Advisor December, 2017 Copyright by Jeffrey M. Gicklhorn 2017 All Rights Reserved UNIVERSITY THE GRADUATE SCHOOL OF NEVADA RENO We recommend that the thesis prepared under our supervision by JEFFREY M. GICKLHORN entitled Succession in a post-fire world: Bunchgrass seedling dynamics after wildfire in sagebrush steppe ecosystems be accepted in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Beth A., Committee Member Paul J., Graduate School Representative David W., Dean, Graduate School December, 2017 i ABSTRACT: Native plant communities experience a constant cycle of disturbance and recovery, and many disturbance regimes are expected to increase in frequency and severity with global change.
Altered disturbance regimes can lead to drastic changes in plant community structure and shifts to alternate states. Ecosystem restoration plays a key role in attempting to return those communities to the appropriate successional trajectory. The Great Basin ecoregion of North America has experienced increasing frequency and size of wildfires coupled with increasing non-native annual grass establishment and widespread domestic livestock grazing. Native bunchgrasses are commonly seeded as restoration treatments after wildfire to stabilize soils and limit annual grass establishment; however, seedings often fail.
Appropriate post-fire livestock management plays an essential role in increasing long-term restoration treatment efficacy. The first chapter examines changing post-fire plant community dynamics over time in the absence of disturbance over two years on two seeded Wyoming big sagebrush sites. Plant community dynamics examined included community composition by functional group, bunchgrass spatial relationships, and factors affecting seedling bunchgrass growth and survival. Seeded functional groups increased with time, suggesting seedings were effective at altering plant community composition.
Bunchgrass spatial relationships initially reflected artificial structure associated with drill seeding; however, spatial patterns shifted over time to reflect plant-plant interactions ii occurring. Bunchgrass seedling growth and survival were negatively affected by increasing neighbor density, and species differed in their responses in year one but not in year two. The second chapter examines the interaction between post-fire plant community structure and timing of initial post fire defoliation over two years on the same sites. I altered plant community structure using removal treatments, and implemented defoliation treatments starting in the first fall after fire.
Seedling removal delayed senescence and decreased bunchgrass cover and density, while adult removal did not have consistent effects. Spring defoliation shortened senescence, and decreased inflorescence production, leaf production, stem length, and total bunchgrass foliar cover. Fall defoliation exhibited mixed effects; however, fall year-two defoliation exhibited fewer negative effects as compared to fall year-one. Seedling removal and spring defoliation interacted to produce the most negative effects, suggesting that defoliating when seedling density is low may be unwise.
General management recommendations include: 1) promoting bunchgrass seedling growing conditions the first year after fire, 2) avoiding spring defoliation all together and delaying fall defoliation until at least the second year after. If initial seedling density is low, delaying livestock further or implement additional restoration treatments. We acknowledge intrinsic differences across sites, and the need for informed and broad management recommendations; however, a site-specific approach is recommended rather than a one-size-fits-all strategy. Lastly, a conservative approach iii to reintroducing livestock is appropriate when one is uncertain about possible negative effects on restored species.
iv ACKNOWLEDGEMENTS: I would like to thank my advisor, Dr. Beth Newingham (USDA-ARS) for her willingness to take me on as a student, push me to excel during my time as a graduate student, and advocate for me when interviewing for professional positions afterward. I would like to thank Beth Leger and Paul Hurtado for great conversations over the last several years and invaluable feedback on my thesis. I would like to thank A.
Munyer for professional assistance and always having an ear when needed; Q. Robleto for their valuable camaraderie and field assistance; D. Veblen for external guidance. I would like to thank the Joint Fire Science Program and the US Fish and Wildlife Service Science Support Partnership for funding my research.
Lastly, I would like to thank my partner, K. Hefty, and family for constant love and support to attend and finish my graduate program with as many amazing experiences as possible. v Table of Contents GENERAL INTRODUCTION. 1 CHAPTER ONE: Community and spatial dynamics of seeded Artemisia tridentata ssp.
wyomingensis shrublands two years after wildfire. 4 Materials and Methods. 8 Field Vegetation Mapping. 9 Community Composition and Species Diversity.
10 Spatial Patterns of Seedling Establishment. 11 Neighborhood Effects on Seedling Growth and Survival. 14 Community Composition and Species Diversity. 14 Spatial Patterns of Seedling Establishment.
15 Neighborhood Effects on Seedling Growth and Survival. 36 CHAPTER TWO: Effects of neighboring plants and defoliation on perennial bunchgrass seedlings after fire in sagebrush communities. 45 Materials and Methods. 49 Vegetation Treatments and Measurements.
55 Tiller Senescence, Growth, Reproduction, and Seedling Survival. 55 Community Foliar Cover and Plant Density. 77 OVERALL SUMMARY AND RECOMENDATIONS. 83 vii LIST OF TABLES: Chapter 1: Table 1.
Seeded native species mixes. Species with * represent locally collected accessions. Seeding rates were not available. ANOVA for foliar cover by functional group, site, and year.
Bold values were statically significant at α = 0.05 and italicized values were significant at α = 0. ANOVA for species diversity (Shannon’s H), and species richness by site and year. Bold values were statically significant at α = 0. Regression results for year one end-of-season seedling size.
Values represent coefficient estimates with standard error in parentheses. Bold values were statically significant at α = 0.05 and italicized values were significant at α = 0. Regression results for seedling survival to year two. Values represent coefficient estimates with standard error in parentheses.
Bold values were statistically significant at α = 0.05 and italicized values were significant at α = 0. Regression results for year two seedling end-of-season size. Values represent coefficient estimates with standard error in parentheses. Bold values were statistically significant at α = 0.05 and italicized values were significant at α = 0.
viii Chapter 2: Table 1. Log-rank comparison significance values for treatment-level Kaplan-Meier curves. Column and row headings denote vegetation treatment combinations, with the upper row denoting neighbor removal and the lower row denoting defoliation treatments. Bolded p-values are significantly different at P < 0.05, and italicized p-values are significantly different at P < 0.
ANOVA table for leaf production, stem length, and flower production by neighbor, defoliation, date, and year. ANOVA table for across-season seedling survival by neighbor, defoliation, and year. ANOVA table for foliar cover and plant density by neighbor, defoliation, plant age class, and year. ix LIST OF FIGURES Chapter 1: Figure 1.
Precipitation for the Coleman (NV) and Saddle Draw (OR) fires (PRISM 2004). Seasons are winter (December of prior year – February), spring (March – May), summer (June – August), and fall (September – November). Mean foliar cover by functional group, site, and year. All functional groups summed represent total foliar cover.
Letters represent statistically significantly different groups for total foliar cover among years and sites, * represent significant differences for a particular functional group within site across years for a particular functional group within site, and † represent significant differences for a particular functional group across sites within year. All comparisons are statistically significant at α = 0. Symbols are only shown on the group with a higher mean but represent the appropriately paired group. A) Shannon’s diversity and B) species richness by site and year.
Points represent fitted model estimates with standard errors. Percentage of plots exhibiting spatial patterns by lag distance in year one and year two for A) seedling bunchgrasses only, and B) adult effects on seedlings. Positive values signify spatial aggregation, negative values signify spatial dispersion for any given lag distance, and values of 0 signify complete spatial randomness for a given lag distance. If both positive and negative values are exhibited at a particular lag distance, x the combination represents a ratio of spatial aggregation to dispersion for that lag distance.
Year one end-of-season seedling size in percent cover by species as a function of neighborhood density within 10cm of seedling. Points represent fitted model estimates and error bars represent 95% confidence intervals. Probability of seedling survival from year one to year two by species as a function of neighborhood density within 10cm of seedling. Points represent fitted model estimates and error bars represent 95% confidence intervals.
Year two end-of-season seedling size by species as a function of year one end- of-season size and year two neighbor density within 10cm. Points represent fitted model estimates and error bars represent 95% confidence intervals. Precipitation for the Coleman (NV) and Saddle Draw (OR) fires (PRISM 2004). Seasons are winter (December of prior year – February), spring (March – May), summer (June – August), and fall (September – November).
Kaplan-Meier curves showing the percent of seedling tillers actively growing as a function of neighbor removal, defoliation, year, and date. * represent significant differences for defoliation treatments as compared to no defoliation within neighbor xi treatment, and † represent significant differences for defoliation treatments relative to no neighbor removal with the same defoliation treatment. Number of actively growing leaves per tiller as a function of neighbor, defoliation, year, and date. Error bars represent 95% confidence intervals for each sample date.
* represent significant differences for defoliation treatments as compared to no defoliation within the same neighbor treatment, and † represent significant differences for defoliation treatments relative to no neighbor removal with the same defoliation treatment. Tiller stem length as a function of neighbor, defoliation, year, and date. Error bars represent 95% confidence intervals for each sample date. * represent significant differences for defoliation treatments as compared to no defoliation within the same neighbor treatment, and † represent significant differences for defoliation treatments relative to no neighbor removal with the same defoliation treatment.
Percent of tillers with inflorescences as a function of neighbor, defoliation, year, and date. Error bars represent 95% confidence intervals for each sample date. * represent significant differences for defoliation treatments as compared to no defoliation within the same neighbor treatment, and † represent significant differences for defoliation treatments relative to no neighbor removal with the same defoliation treatment. A) Percent foliar cover and B) plant density as a function of treatment type, age class, and year.
Bars represent model perimeter estimates and error bars represent 95% confidence intervals. Column headings denote neighbor removal (upper row) and defoliation (lower row) treatments. Dark gray bars represent adult cover and light gray bars represent seedling cover. 6 A has a dashed line at 20% foliar cover to denote the suggested management benchmark for reintroduction of livestock grazing after fire.
1 GENERAL INTRODUCTION Native plant communities experience a constant cycle of disturbance and recovery, and many disturbance regimes are expected to increase in frequency and severity with global change (Spracklen et al. Altered disturbance regimes can lead to drastic changes in plant community structure and composition and possible shifts to alternate dominant species. Ecosystem restoration plays a key role in attempting to return those communities to the correct successional trajectory after disturbance and reestablishing communities resilient to future disturbances. Additionally, appropriate post-disturbance management is essential to increase the efficacy of these restoration treatments by allowing seeded species to establish and limit the establishment of non- native species.
The Great Basin ecoregion of North America has experienced drastic shifts in wildfire frequency and size. This shift coupled with increasing presence of non-native annual grasses has created a grass-fire feedback loop, leading to the loss of native sagebrush steppe communities and further increasing fire and annual grass invasion.