Doctor of Philosophy / Department of Division of Biology / John Blair / The feasibility of restoration, which traditionally targets historical conditions, is questionable in the context of global change. To address this, my dissertation investigated (Chapter 2) the patterns of restoration establishment along a chronosequence of restored prairies with respect to nearby remnant prairies, (Chapters 3-4) responses of plant communities in restorations initiated using different methods (levels of species richness and sowing density) to drought, which is projected to increase in frequency, and (Chapters 5-6) the effects of propagule source and variation (mixing among sources) on restoration establishment and the generality of restoration outcomes across variable environments using reciprocal common gardens of multi-species restoration seedings. Chapter 2, published in Restoration Ecology, showed that restoration led to the recovery of desirable characteristics within several years, but restorations utilizing primarily fall-collected seeds likely diminished the representation of early phenology species, so biodiversity may be further enhanced by including early phenology species in seeding mixes. Chapters 3 and 4, published respectively in Ecological Applications and Applied Vegetation Science, examined the establishment of native plant communities after seeding and their responses to experimentally imposed drought. Both high seed mixture richness and high density seeding resulted in greater establishment of native, seeded species compared to low richness and low density treatments, and exotic species were less prevalent in high richness and high density treatments. However, we found little evidence of differential drought resistance, recovery, and resilience among treatments. This result coupled with increases in exotic species following drought suggest that other forms of active management may be needed to produce restored plant communities that are robust to climate change. Chapter 5 (published in Ecosphere) iii and Chapter 6 found that seed source affects individual species establishment, community structure, and productivity. However, there was no consistent advantage for any source, including local sources, across sites or species. This suggests that source effects on single species or effects observed at single locations should not be broadly generalized. Together, this dissertation shows that restoration can recover many characteristics of native prairies and that manipulation of seeding practices (seed mixture richness, seeding density, seed source) influence grassland establishment in terms of productivity, community structure, invasion, and the abundance and survival of individual species.

Doctor of Philosophy / Department of Biology / David C. Hartnett / Grazing and drought are instrumental in the development and maintenance of perennial grasslands. In this research I tested the belowground bud bank contribution to tallgrass prairie resistance and resilience when perturbed by grazing and drought. First, I tested the bud bank role in vegetation response to and recovery from severe drought (Chapter 2). I compared above- and belowground responses of experimentally droughted plots to ambient controls and irrigated plots during two years of severe drought and two years of recovery. I found that although aboveground net primary productivity declined 30-60% during drought, bud bank density and demography were insensitive to drought. These results suggest that grassland resistance and resilience when perturbed by drought may be mediated by stability of belowground bud banks. Second, I investigated vegetation and soil nutrient legacies following release from long-term grazing (Chapter 3). I documented a relatively rapid shift in aboveground vegetation within four years of grazer exclusion, with productivity, stem density, and diversity becoming relatively more similar to ungrazed than grazed prairie. The density and composition of the belowground bud bank and soil seed bank shifted more slowly, remaining more similar to grazed than ungrazed prairie. Responses of soil nutrients to removal of grazers varied, and in some cases was affected by recent fire history. These results demonstrate the contribution of belowground propagules to the maintenance of a diverse plant community both during grazing and after grazers are removed. Finally, I examined short-term vegetation responses to both drought and grazing (Chapter 4). Despite extreme drought and simulated grazing that reduced productivity and increased mortality of individual stems, the dominant C4 grasses maintained a stable bud bank. Aboveground net primary productivity and bud bank density of sedges and forbs, however, were reduced by both drought and grazing. This differential response of species to extreme drought and grazing led to shifts in community composition and species diversity over one growing season. Across drought and grazing treatments, live rhizome biomass was highly correlated with bud bank density and may be a useful, more easily measured index of bud bank density.

Doctor of Philosophy / Department of Agronomy / Charles W. Rice / Increasing atmospheric CO2 concentrations and other greenhouse gases have been linked to global climate change. Soil organic C (SOC) sequestration in both agricultural and native ecosystems is a plausible option to mitigate increasing atmospheric CO2 in the short term. Laboratory and field studies were conducted to (1) understand the influence of soil water content on the temperature response of SOC mineralization (2) investigate burn and nutrient amendment effects on biogeochemical properties of tallgrass prairie and (3) assess perennial and annual plant management practices on biophysical controls on SOC dynamics. The laboratory study was conducted using soils collected from an agricultural field, currently planted to corn (C4 crop), but previously planted to small grain (C3) crops. The changes in cultivated crops resulted in a δ¹³C isotopic signature that was useful in distinguishing older from younger soil derived CO2-C during SOC mineralization. Soils were incubated at 15, 25 and 35 oC, under soil water potentials of -1, -0.03 and -0.01 MPa. Soil water content influenced the effect of temperature on SOC mineralization. The impact of soil water on temperature effect on SOC mineralization was greater under wetter soil conditions. Both young and older SOC were temperature sensitive, but SOC loss depended on the magnitude of temperature change, soil water content and experiment duration. Microbial biomass was reduced with increasing soil water content. The first field experiment investigated burn and nutrient amendment effects on soil OC in a tallgrass prairie ecosystem. The main plots were burned (B) and unburned (UB) tallgrass prairie and split plots were nutrient amendments (N, P or N+P including controls). Vegetation was significantly altered by burning and nutrient amendment. Treatment effects on either TN or SOC were depth-specific with no impact at the cumulative 0-30 cm depth. The P amendment increased microbial biomass at 0-5 cm which was higher in unburned than burned. However, at 5-15 cm depth N amendment increased microbial biomass which was higher in burned than unburned. In conclusion, soil OC in both burned and unburned tallgrass prairie may have a similar trajectory however; the belowground dynamics of the burned and unburned tallgrass prairie are apparently different. Another field experiment assessed SOC dynamics under perennial and annual plant management practices. The main plots were grain sorghum (Sorghum bicolor) planted in no-tillage (NT) or continuous tillage (CT), and replanted native prairie grass, (Andropogon gerardii) (RP). The spit plots were phosphorus (+P) and control without P (-P). The P amendment was used to repress arbuscular mycorrhizal fungi (AMF), known to influence soil aggregation. The macroaggregate >250 µm, SOC and TN were higher in RP and NT than CT. The relative abundances of AMF and saprophytic fungi were greater with less soil disturbance in RP and NT than in CT. Therefore, less soil disturbance in RP and NT increased AMF and fungal biomasses. The higher relative abundances of AMF and fungi with less soil disturbance increased macroaggregate formation in RP and NT, which resulted in higher SOC sequestration in RP and NT than CT.

Snake populations in the North American tallgrass prairie appear to be declining, yet data unavailability impedes the development of enhanced ecological understanding of snake species-habitat relationships and also hinders snake conservation efforts. This study addresses both issues for the snakes of Illinois in two steps. In a two-year mark-recapture study at twenty-two sites within six northern Illinois prairie preserves, I investigated snake species-habitat relationships using habitat variables at three scales: microhabitat (< 100 m), landscape (1 - 10 km), and regional (> 10 km). A total of 120 snakes representing seven species was captured using drift fence arrays associated with funnel traps and sheet metal cover. The low numbers and diversity of snakes captured, when compared to historic evidence, indicate that Illinois snake populations have declined. At the microhabitat scale, non-metric multidimensional scaling and Mantel tests revealed a relationship between snake species composition and elevation. At the landscape-scale, snake species composition varied along an agricultural-urban cover gradient. Classification and regression trees and maximum entropy models (Maxent) were used to identify the scales at which snake species-habitat relationships were strongest. Six of seven regression trees for individual snakes species contained habitat variables at the landscape scale. Important landscape characteristics included patch size, isolation, and land cover, metrics that strongly covary with habitat loss. Microhabitat features only appeared in the regression trees of two species and in three Maxent models. This study indicates that habitat loss has shaped the current distribution of snake species in Illinois's remnant prairies and that snake conservation efforts should emphasize the landscape-scale. Finally, I developed a risk ranking system based on natural and life history characteristics to assess the conservation status of Illinois's 38 snake species. Cluster analysis identified eight groups of snakes, similar in terms of risk factors, with high risk species sharing characteristics such as large body size, long life span, limited habitat breadth, and a high anthropogenic threat ranking. Here, I emphasize the need for basic demographic studies on snakes and suggest that ranking systems be used with population data (when available) and expert opinion to identify snake species of conservation concern in other regions. / Dissertation

Prescribed burning is an important management tool for the restoration and maintenance of tallgrass prairies. To improve fire behaviour prediction in tallgrass prairies, I assessed three different aspects of fire behaviour - heat of combustion, fuel load and rate of spread. Heat of combustion was found to vary amongst certain tallgrass species but the relatively small differences in means is unlikely to contribute significantly to fire behaviour. Average fuel loads in Ontario tallgrass prairie sites were found to be higher than current default value used in fire behaviour prediction. Three rapid fuel load assessment techniques were tested. Finally, the predictions of three fire behaviour prediction systems - the FBP System, BehavePlus and an Australian grassfire spread model, were compared with actual fire behaviour observations. The FBP System was found to perform poorly while both BehavePlus and the Australian model exhibited relatively strong relationships between observed and predicted rates of spread.

Prescribed burning is an important management tool for the restoration and maintenance of tallgrass prairies. To improve fire behaviour prediction in tallgrass prairies, I assessed three different aspects of fire behaviour - heat of combustion, fuel load and rate of spread. Heat of combustion was found to vary amongst certain tallgrass species but the relatively small differences in means is unlikely to contribute significantly to fire behaviour. Average fuel loads in Ontario tallgrass prairie sites were found to be higher than current default value used in fire behaviour prediction. Three rapid fuel load assessment techniques were tested. Finally, the predictions of three fire behaviour prediction systems - the FBP System, BehavePlus and an Australian grassfire spread model, were compared with actual fire behaviour observations. The FBP System was found to perform poorly while both BehavePlus and the Australian model exhibited relatively strong relationships between observed and predicted rates of spread.

Master of Science / Department of Animal Sciences and Industry / K C Olson / We evaluated effects of annual, prescribed burning on vigor of the noxious weed, sericea lespedeza (Lespedeza cuneata; SL) in native tallgrass prairie over a 4-yr period. We hypothesized that annual prescribed burning conducted during the growing season would selectively pressure SL, whereas locally-conventional, dormant-season prescribed burning would have no effect on SL. A 50-ha native tallgrass pasture infested with SL (initial basal frequency = 2 ± 1.3%, initial aerial frequency = 36 ± 3.4%) was used for our study. It was divided along watershed boundaries into 9 fire-management units (5 ± 2.6 ha) for this experiment. Burn units were assigned randomly to 1 of 3 prescribed-burning times (n = 3 / treatment): early spring (1 April; EARLY), mid-summer (1 August; MID), or late summer (1 September; LATE). Forage biomass, SL aerial frequency, SL stem length, SL seed production, soil cover, and plant species composition were measured along single, permanent 100-m transects in each burn unit. Treatment and measurement date influenced forage biomass and SL stem length (treatment × time). Forage biomass was not different (P ≥ 0.43) between treatments on 17 July; however, forage biomass was greater (P < 0.01) in EARLY than MID and greater in MID than LATE on 10 October. Maximum stem length of SL was less (P ≤ 0.02) in MID and LATE than in EARLY on 17 July and on 10 October. Aerial frequency of SL was least (main effect – P < 0.01) in LATE, intermediate in MID, and greatest in EARLY, whereas basal frequency of SL was less (P < 0.01) in MID and LATE compared with EARLY. Whole-plant dry weight and seed production of SL at dormancy were greatly diminished (P ≤ 0.02) in MID and LATE compared with EARLY. Occurrence of bare soil, litter cover, and total basal plant cover were not different (P ≥ 0.21) between treatments. Similarly, basal cover of grasses, forbs, and shrubs were not different (P ≥ 0.24) between treatments. We interpreted these data to indicate that annual prescribed burning during the growing season had strong suppressive effects on SL compared to locally-conventional, early-season prescribed burning and produced no apparent detrimental effects on soil cover or non-target plant species. Post-fire regrowth was sufficient to prevent erosion and soil-moisture loss during the subsequent dormant season and would have allowed light to moderate grazing during the ensuing winter. Key words: Lespedeza cuneata, prescribed fire, range improvement

Master of Science / Department of Biology / Brett K. Sandercock / Grassland birds are declining throughout their native range. The Flint Hills of eastern Kansas and Oklahoma contain large tracts of tallgrass prairie, but intensification of agricultural practices may be contributing to ongoing population declines. Common rangeland management practices include annual burning coupled with heavy grazing by cattle. This system, known as intensive early stocking and burning, promotes homogeneous utilization of forage by cattle but may not provide habitat for some grassland bird species. Patch-burn grazing is an alternative management system that aims to restore heterogeneity on rangelands by recreating the fire-grazing interaction that would have historically occurred throughout the Great Plains. From 2011-2013, we examined responses of grassland birds to traditional rangeland management and patch-burn grazing by conducting vegetation surveys, line transect surveys, and nest monitoring on privately-owned pastures in Chase County and Greenwood County, Kansas. Vegetative heterogeneity was higher on patch-burned pastures, with unburned patches having higher visual obstruction and less bare ground. Densities of grassland birds differed by species and among habitat strata. Unburned patches on patch-burned pastures were associated with increased densities of Dickcissels (Spiza americana), Eastern Meadowlarks (Sturnella magna) and Grasshopper Sparrows (Ammodramus savannarum). Henslow’s Sparrows (A. henslowii) were only detected on patch-burned pastures. Nest survival of grassland songbirds was similar among management systems but varied by year. Probability of nest parasitism by Brown-headed Cowbirds (Molothrus ater) varied among years and between treatments for Dickcissels and Grasshopper Sparrows, with overall lower rates on burned areas and during drought years. For Dickcissels and Grasshopper Sparrows, there was a significant reduction in host clutch size between parasitized versus unparasitized nests. Overall, nest survival of grassland songbirds in managed rangelands was low. Patch-burn grazing improved rangeland conditions and provided habitat for more species of birds, but did not increase nest survival. Drought conditions in 2012 and 2013 may have influenced the results of this study, as many landowners were unable to burn as planned. Further study is needed to determine underlying factors driving variation in nest success and parasitism rates for grassland birds, particularly on private lands which make up the vast majority of remnant tallgrass prairies.

Master of Science / Department of Civil Engineering / David R. Steward / Konza Prairie is one of the few areas in the United States were natural landscape of the area is still intact. Human action on changing the landscapes in this area is limited and much of the land remains as native grassland. In spite of its natural existence, this area is not completely isolated from the rest of the world. Changes that are taking place in climate will eventually have the same effect to this region as well as other human populated areas. Increase in carbon concentration in the air has resulted to increase in temperature, this increase in temperature increases the evaporation from the sea, oceans and the ice capes. As the atmospheric water vapor changes the precipitation pattern also change. Changes in precipitation due to climate change will result to change in hydrology and hydraulics of the streams and groundwater flow regime. Precipitation provides surface runoff and groundwater infiltration, which recharge the cracked limestone aquifer present in the Konza area. The infiltration water moves trough the cracked rocks and eventually reach the creeks such as Kings Creek and flow to the Kansas River. Increase in precipitation will result to increase in surface runoffs and more groundwater recharge. Decrease in precipitation will result to decrease in both surface and groundwater. To examine changes in groundwater elevation as recharge change in Konza, a groundwater model was developed based on erosion impact calculator (EPIC) ecological model and SLIT groundwater model. EPIC model estimates the deep percolation (recharge) as 12% and total runoff to about 24% of the annual average precipitation. The annual average recharge values from EPIC were used in SPLIT to simulate results for the groundwater elevation at Konza prairie. Field wells elevation were use to calibrate the SPLIT results. By estimating the hydraulic permeability value to 0.546m/d the field well measurements and SPLIT simulated groundwater elevation results provide a good match. After calibration max and min recharge together with a 5-years moving average were used to examine the changes in groundwater elevation as recharge changes. Future study intends to use the calibrated Konza groundwater model and the forecasted climate data to simulate result for groundwater elevation as climate changes.

Master of Science / Department of Biological & Agricultural Engineering / Stacy L. Hutchinson / Urban stormwater is one of the leading causes of water quality impairment and stream channel degradation in the United States. In an effort to address the negative effects of stormwater runoff on receiving aquatic systems, Best Management Practices for stormwater, including ecologically-designed stormwater systems, are becoming more common across the urban landscape. Throughout eastern Kansas and the rest of the Midwestern United States, prairie grasses are beginning to receive attention for their potential to enhance infiltration within these systems. However, the function of vegetated stormwater systems and the influence of factors such as vegetation age on infiltration and system performance are not well understood because monitoring data for these systems is limited. When performance data is collected, it often pertains only to the hydraulic and water quality aspects of the system but neglects any assessment of the integrity of the ecosystem functions on which the system's performance is dependent. The objective of this study was to address the need for an assessment tool that considers the ecological integrity, or health, of ecologically-designed stormwater systems, as well as to fill the gap in the literature regarding the function of ecologically-designed stormwater systems in the tallgrass prairie region. Since many of the eco-based stormwater practices in the region rely upon the establishment of native prairie grasses to enhance infiltration on the site, the specific focus of this study was to gain a better understanding of infiltration processes in ecologically-designed systems and the extent of our ability to regain these processes through prairie restoration in previously disturbed urban sites. To address these objectives, two stormwater systems at different stages of vegetative maturity were examined. In general, ecosystem health scores were higher for the more mature system and could be used to guide future management decisions at both sites. Results from the hydraulic analysis indicate the function of the system may improve over the course of the growing season, but statistical relationships between system age and infiltration rate could not be established.