Susan E. Meyer

WHAT MAKES GREAT BASIN SAGEBRUSH ECOSYSTEMS INVASIBLE BY BROMUS TECTORUM

Ecosystem susceptibility to invasion by nonnative species is poorly understood, but evidence is increasing that spatial and temporal variability in resources has large-scale effects. We conducted a study in Artemisia tridentata ecosystems at two Great Basin locations examining differences in resource availability and invasibility of Bromus tectorum over elevation gradients and in response to direct and interacting effects of removal of perennial herbaceous vegetation and fire. We monitored environmental conditions, soil variables, and B. tectorum establishment and reproduction over two years. Soil water (measured as the number of days soil matric potential was .� 1.5 MPa) and nitrate availability (measured as micromoles of NO3 � sorbed to resin capsules per day in the ground) decreased with decreasing elevation. Lower-elevation sites had greater annual variability in soil water availability than upper-elevation sites did. Soil nitrate levels were highest at all elevations when soils were wettest; nitrate availability was not more variable at lower elevations. Removal of herbaceous perennials increased soil water and nitrate availability, but burning without removal had only minor effects. Bromus tectorum had low establishment, biomass, and seed production on high-elevation sites and on a mid-elevation site during a cold, short, growing season probably due to ecophysiological limitations resulting from cold temperatures. Establishment, biomass, and seed production were variable at low elevations and best explained by soil characteristics and spatial and temporal variation in soil water. Removal and fire had minor effects on emergence and survival, but biomass and seed production increased two to three times following removal, two to six times after burning, and 10-30 times following removal and burning. Our data indicate that invasibility varies across elevation gradients and appears to be closely related to temperature at higher elevations and soil water availability at lower elevations. High variability in soil water and lower average perennial herbaceous cover may increase invasion potential at lower elevations. Soil water and nitrate availability increase following either fire or removal, but on intact sites native perennials typically increase following fire, limiting B. tectorum growth and reproduction. Following resource fluctuations, invasibility is lowest on sites with relatively high cover of perennial herbaceous species (i.e., sites in high ecological condition).

Ecological genetics of seed germination regulation in Bromus tectorum L.

Abstract Regulation of seed germination phenology is an important aspect of the life history strategy of invading annual plant species. In the obligately selfing winter annual grass Bromus tectorum, seeds are at least conditionally dormant at dispersal in early summer and lose dormancy through dry-afterripening. Patterns of germination response at dispersal vary among populations and sometimes across years within populations. To assess the relative contribution of genotype and maturation environment to this variation, we grew progeny of ten parental lines from each of six contrasting populations in a common greenhouse environment. We then tested the germination responses of recently harvested seeds of the putative full-sib progeny at five incubation temperatures. Significant germination response differences among populations were observed in greenhouse cultivation, and major differences among full-sib families were evident for some populations and traits. Among-population variation accounted for over 90% of the variance in each trait, while within-family variance accounted for 1% or less. Germination responses of greenhouse-grown progeny were positively correlated with the responses of wild-collected seeds, but there was a tendency for lowered dormancy at higher incubation temperatures. This tendency was more marked in populations from cold desert, foothill, and plains habitats, suggesting a genotype-maturation environment interaction. Differences among populations in the amount of among-family variance were more evident at lower incubation temperatures, while among-family variance was more uniformly low at summer incubation temperatures. Populations from predictable extreme environments (subalpine meadow and warm desert margin) showed significantly less variation among families than populations from less predictable cold desert, foothill, and plains environments. Low among-family variance was not specifically associated with small population size or marginality of habitat, as small marginal populations from unpredictable environments showed variance as high as that of large populations. In populations with high among-family variance for germination traits, germination responses tended to be correlated across incubation temperatures, making it possible to characterize families in terms of their general dormancy status. The results indicate that seed germination regulation in this species is probably under strong genetic control, and that habitats with temporally varying selection are occupied by populations that tend to be more polymorphic in terms of their germination response patterns.

Ecological aspects of seed dormancy loss

Advances in seed biology include progress in understanding the ecological significance of seed dormancy mechanisms. This knowledge is being used to make more accurate predictions of germination timing in the field. For several wild species whose seedlings establish in spring, seed populations show relevant variation that can be correlated with habitat conditions. Populations from severe winter sites, where the major risk to seedlings is frost, tend to have long chilling requirements or to germinate very slowly at low temperatures. Populations from warmer sites, where the major risk is drought, are non-dormant and germinate very rapidly under these same conditions. Seed populations from intermediate sites exhibit variation in dormancy levels, both among and within plants, which spreads germination across a considerable time period. For grasses that undergo dry after-ripening, seed dormancy loss can be successfully modelled using hydrothermal time. Dormancy loss for a seed population is associated with a progressive downward shift in the mean base water potential, i.e., the water potential below which half of the seeds will not germinate. Other parameters (hydrothermal time requirement, base temperature and standard deviation of base water potentials) tend to be constant through time. Simulation models for predicting dormancy loss in the field can be created by combining measurements of seed zone temperatures with equations that describe changes in mean base water potential as a function of temperature. Successful validation of these and other models demonstrates that equations based on laboratory data can be used to predict dormancy loss under widely fluctuating field conditions. Future progress may allow prediction of germination timing based on knowledge of intrinsic dormancy characteristics of a seed population and long-term weather patterns in the field.

Seed germination regulation in Bromus tectorum (Poaceae) and its ecological significance

Bromus tectorum is a winter annual grass that has become extensively naturalized in western North America. Its seeds are usually at least conditionally dormant at dispersal and lose dormancy through dry afterripening. Germination response to temperature for recently harvested seeds and rate of change in germination response during afterripening were examined for collections from 21 western North American populations representing a wide array of habitats. Analysis of variance showed highly significant among-population differences in germination response variables. Principal components analysis of 20 germination variables revealed groups of populations that could be characterized by distinct response syndromes. Degree of dormancy at summer temperatures in recently harvested seeds as well as rate of dormancy loss during dry storage could be related to the risk of premature summer germination in different habitats. Mojave Desert populations showed the most clearly differentiated response. Populations from Intermountain desert and foothill habitats showed intermediate responses and did not form distinct groups. Montane populations showed the widest variation. Fully afterripened seeds from all populations were nondormant and could germinate quickly across a wide temperature range. These results demonstrate the existence of adaptively significant variation in germination response. Such variation probably represents the beginning of genetic differentiation as a result of selection among and within founder populations. Lack of a consistent relationship with habitat reflects the stochastic nature of colonization and the fact that diverse germination strategies may permit persistence, especially in less extreme habitats.

Using hydrothermal time concepts to model seed germination response to temperature dormancy loss and priming effects in Elymus elymoides

Hydrothermal time (HTT) describes progress toward seed germination under various combinations of incubation water potential ( ) and temperature ( T ). To examine changes in HTT parameters during dormancy loss, seeds from two populations of the bunchgrass Elymus elymoides were incubated under seven temperature regimes following dry storage at 10, 20 and 30°C for intervals from 0 to 16 weeks. Fully after-ripened seeds were primed for 1 week at a range of s. Data on germination rate during priming were used to obtain a HTT equation for each seed population, while data obtained following transfer to water were used to calculate HTT accumulation during priming. HTT equations accurately predicted germination time course curves if mean base water potential, b ( 50 ), was allowed to vary with temperature. b ( 50 ) values increased linearly with temperature, explaining why germination rate does not increase with temperature in this species. b ( 50 ) showed a linear decrease as a function of thermal time in storage. Slopes for the T × b ( 50 ) relationship did not change during after-ripening. This thermal after-ripening time model was characterized by a single base temperature and a constant slope across temperatures for each collection. Because the difference between initial and final b ( 50 )s was uniform across tempera-tures, the thermal after-ripening requirement was also a constant. When seeds were primed for 1 week at −4 to −20 MPa, accumulation of HTT was a uniform 20% of the total HTT requirement. When primed at 0 to −4 MPa, HTT accumulation decreased linearly with decreasing priming potential, and a hydrothermal priming time model using a constant minimum priming potential adequately described priming effects. Use of these simple HTT relationships will facilitate modelling of germination phenology in the field.

A hydrothermal time model of seed after-ripening in Bromus tectorum L.

Bromus tectorum L. is an invasive winter annual grass with seeds that lose dormancy through the process of dry after-ripening. This paper proposes a model for after-ripening of B. tectorum seeds based on the concept of hydrothermal time. Seed germination time course curves are modelled using five parameters: a hydrothermal time constant, the fraction of viable seeds in the population, base temperature, mean base water potential and the standard deviation of base water potentials in the population. It is considered that only mean base water potential varies as a function of storage duration and incubation temperature following after-ripening. All other parameters are held constant throughout after-ripening and at all incubation temperatures. Data for model development are from seed germination studies carried out at four water potentials (0, −0.5, −1.0 and −1.5 MPa) at each of two constant incubation temperatures (15 and 25°C) following different storage intervals including recently harvested, partially after-ripened (stored for 4, 9 or 16 weeks at 20°C) and fully after-ripened (stored for 14 weeks at 40°C). The model was fitted using a repeated probit regression method, and for the two seed populations studied gave R 2 values of 0.898 and 0.829. Germination time course curves predicted by the model generally had a good fit when compared with observed curves at the incubation temperature/water potential treatment combinations for different after-ripening intervals. Changes in germination time course curves during after-ripening of B. tectorum can largely be explained by decreases in the mean base water potential. The simplicity and good fit of the model give it considerable potential for extension to simulation of after-ripening under field conditions.

GERMINATION RESPONSE OF ARTEMISIA TRIDENTATA (ASTERACEAE) TO LIGHT AND CHILL: PATTERNS OF BETWEEN-POPULATION VARIATION

Artemisia tridentata Nutt. is the dominant shrub over much of semiarid North America. It consists of several subspecies and includes ecotypes that occur over a range of habitats from warm desert fringes to montane forest and meadow communities. Fifteen seed collections representing the three common subspecies and a spectrum of habitats were subjected to a series of laboratory light and chill treatments. Recently harvested seeds were mostly nondormant at 15 C but required light for full germination. Removal of the pericarp, afterripening in dry storage, and short chill treatments all resulted in a reduction in light requirement. When examined on a by-collection basis, seed germination response variables were significantly correlated with each other and with mean January temperature at the seed collection site, an index of winter severity. Collections from colder sites were 16%-36% dormant in the light at 15 C and nearly 100% light-requiring, while collections from warm desert fringes were nondormant in light and only 50%-70% light-requiring. Cold winter collections required up to 98 d to germinate to 50% at 1 C in light, whereas warm winter collections germinated to 50% in as little as 16 d. There was no clear relationship between germination behavior and subspecific identity. The observed climate-correlated variation in germination response to light and chill appears to be of adaptive significance, but a genetic basis for patterns of infraspecific variation in germination response in this species has not yet been demonstrated.

Genetic variation and local adaptation at a cheatgrass (Bromus tectorum) invasion edge in western Nevada

Cheatgrass (Bromus tectorum) is an invasive weed in western North America found primarily growing at elevations less than 2200 m. We asked whether cheatgrass is capable of becoming adapted to a marginal habitat, by investigating a population at a high elevation invasion edge. We used a combination of methods, including reciprocal field transplants, controlled environment studies and molecular analysis. High levels of SSR gene diversity (0.50 vs. 0.43) and comparable variation in phenotypic traits were observed at both the invasion edge and a low elevation, high‐density population. Three heterozygotes were observed in the edge population, which is unusual in this predominantly self‐pollinating plant. Plants from high elevations germinated more slowly in a growth chamber and had slower seedling growth rates. Survivorship was low at the edge (13%), compared with the low elevation site (55%), but surviving plants were of similar size and had equivalent reproductive output. Seed size positively affected survival and plant performance in the field and this trait was inherited. Emergence timing affected survival at the low elevation site and germination timing was also inherited. Local adaptation was seen in the low, rather than in the high elevation site, because of differential survival. While there was no evidence for local adaptation to the high elevation site observed in the field, family level and genotype‐level differences in traits that affected field performance, high genetic diversity at the invasion edge, and evidence of outcrossing in this highly selfing species indicates that the potential for adaptation to a marginal habitat exists within this population.

A hydrothermal after-ripening time model for seed dormancy loss in Bromus tectorum L.

After-ripening, the loss of dormancy under dry conditions, is associated with a decrease in mean base water potential for germination of Bromus tectorum L. seeds. After-ripening rate is a linear function of temperature above a base temperature, so that dormancy loss can be quantified using a thermal after-ripening time (TAR) model. To incorporate storage water potential into TAR, we created a hydrothermal after-ripening time (HTAR) model. Seeds from two B. tectorum populations were stored under controlled temperatures (20 or 30 °C) and water potentials (−400 to −40 MPa). Subsamples were periodically removed from each storage treatment and incubated at 15 or 25 °C to determine germination time courses. Dormancy status (mean base water potential) was calculated from each time course using hydrothermal time equations developed for each seed collection. Seeds stored at −400 MPa did not after-ripen. At water potentials from −400 to −150 MPa, the rate of after-ripening increased approximately linearly with increasing water potential. Between −150 and −80 MPa, there was no further increase in after-ripening rate, while at −40 MPa seeds did not after-ripen and showed loss of vigour. These results suggest that the concept of critical water potential thresholds, previously shown to be associated with metabolic activity and desiccation damage in partially hydrated seeds, is also relevant to the process of after-ripening. The HTAR model generally improved field predictions of dormancy loss when the soil was very dry. Reduced after-ripening rate under such conditions provides an ecologically relevant explanation of how seeds prolong dormancy at high summer soil temperatures.

Big sagebrush germination patterns: subspecies and population differences.

Habitat-correlated differences in laboratory germination response under autumn (15 degrees C) and winter (1 degree C) temperature regimes were examined for 69 big sagebrush (Artemisia tridentata Nutt., Asteraceae) seed collections from a range of habitats in 7 western states. Mountain big sagebrush (ssp. vaseyana) exhibited the widest variation in dormant seed percentage and termination rate at 15 degrees C. Collections from severe winter sites had larger dormant seed fractions and slower germination rates than collections from mild winter sites. Basin big sagebrush (ssp. tridentata) and Wyoming big sagebrush (ssp. wyomingensis) collections were largely non-dormant and germinated quickly at 15 degrees C regardless of collection site winter climate. At 1 degree C, number of days to 50% of total germination was negatively correlated with collections site mean January temperature for all 3 subspecies. Collections from severe winter sites required up to 113 days to germinate to 50% at 1 degree C, while collections from mild winter sites required as few as 6 days. Habitat-correlated variation in germination response appears to be of adaptive significance. Dormancy and slow germination at 15 degrees C may prevent germination during autumn storms in the mountains, while delayed germination at continuous 1 degree C may prevent precocious germination under snowpack. In contrast, at mild winter sites, winter germination is promoted and probably affords the best chance for seedling survival. Between-population variation in germination strategy should be considered when artificially seeding this species.