Project Number: 2070-22000-004-03
Start Date: Sep 01, 2011
End Date: Aug 31, 2015
The general approach for evaluating spatial and temporal variation in seedling demography as well as ecological processes associated with this variation and potential management solutions are as follows: In fall of each of the three study years we will seed monocultures of three representative species in 1m2 plots using three replicate plots per species per year. Species and years will be randomly assigned to plots. To keep the intensity of disturbance and seed bed conditions comparable among years plots were tilled to a depth of 8 cm and existing vegetation removed approximately one month before planting. Volumetric soil moisture and temperature sensors will be installed at three randomly selected locations within the study area and measurements were made hourly in the 5 cm and 15 cm soil layer. Germination will be measured using the buried bag technique (Abbott and Roundy 2003). For each species and year, 40 bags will be randomly paired with the seeded 1m2 plots and planted in fall at the same time the plots were seeded. Five bags of each species will be pulled approximately every two weeks starting in winter and continuing through spring. At the same time, seedling emergence and death will be tracked weekly on the seeded plots. Individual seedlings within a cohort will be marked with colored toothpicks. We will manipulate soil moisture through irrigation and soil pathogens through fungicide additions to examine effects on seedling establishment. We also will examine variation in emergence probabilities across 10 Elymus elymoides accessions and 10 Agropyron desertorum accessions. We will use the developed procedures of Madsen et al. (2011) to evaluate the effects of seed agglomeration and seed coating technology on seedling establishment. Our statistical estimates will be 95% Bayesian confidence intervals (CIs). Bayesian confidence intervals have a simple interpretation and are well-suited for quantifying survival probabilities and other parameters (Rinella and James 2010). When two 95% confidence intervals do not overlap, the probability is greater than 0.95 that the treatment with the larger-valued interval is larger than the other treatments.