Using Predictive Modeling Tools to Improve Timing of Seeding Treatments

Using Predictive Modeling Tools to Improve Timing of Seeding Treatments

Researchers at BYU used SageSTEP soil moisture and temperature data to develop a tool that predicts germination in the sagebrush steppe and shows how some species are more likely to experience premature germination when sown in the fall a step toward better rangeland seeding practices.

Matt Madsen knows that getting seeds to grow in sagebrush-steppe can be tricky. Timing of germination strongly impacts whether or not seeds eventually survive – and timing depends on multiple factors including exposure to pathogens, available nutrients, soil moisture, temperature, light, and herbivory. Managers could leverage these factors to make seeding efforts more successful, but tracking of seed germination in the field is difficult and time consuming – and produces limited useful information anyway, since short-term studies can’t take into account high annual variability in weather. 

Seeding treatments in the sagebrush-steppe typically occur in the fall, with the expectation that seeds will remain dormant over winter and germinate in spring, said Madsen, assistant professor of Plant and Wildlife Sciences at Brigham Young University. But planting too early in the year can result in early germination and mortality over winter. Understanding appropriate seeding dates could prevent premature germination and subsequent winter mortality – and ultimately improve the success of restoration projects.

Understanding germination characteristics of individual species may help guide land managers in their restoration efforts. For example, planting sagebrush (Artemisia tridentata) in mid-October is late enough to avoid winter germination on average; but species like bluebunch wheatgrass (Pseudoroegeneria spicata) germinate more quickly, and need to be planted in mid-December to avoid high rates of germination over the winter. To complicate things further, seeding plans carefully developed for one site do not necessarily translate to other sites or years with different soil temperature and moisture regimes. 

To help managers improve the success of their seeding efforts, researchers like Madsen have turned to predictive germination models. These use the natural processes within seeds that regulate germination timing (mostly a function of temperature and moisture for non-dormant seeds). Researchers have learned to predict germination of cool-season species through wet-thermal accumulation models, which predict the rate that seeds will germinate in the field based on soil temperature when soil moisture is above set threshold. They’ve found that wet-thermal accumulation models are fairly accurate at predicting seed germination timing in the field. But even with this model, large amounts of data and processing are needed to develop accurate estimates. 

Now that process can be easier. To make models more usable for managers, Madsen and coauthors created a programmed workbook called Auto-Germ which allows users to process seed germination data and predict germination timing in the field. It helps users create wet thermal accumulation models from laboratory germination trials conducted over a range of temperatures. Auto-Germ gives users an interface to apply the wet-thermal accumulation models to historic field soil moisture and temperature data sets to estimate seed germination timing.

This information can help managers know how planting dates may influence germination and subsequent chances for viability based on growing conditions. Research published in 2018 provides instructions on how to use Auto-Germ. It uses a case study to calculate various germination indices under different constant temperatures on species commonly used for restoration projects in the Great Basin. It also calculates germination timing for six years across ten sites to estimate the planting date required for 50% or more of the simulated population of seeds to germinate in spring when conditions could be more conducive for plant establishment.

Based on their results, Madsen anticipates that Auto-Germ will be applicable to non-dormant seeds of most species. Both land managers and researchers could benefit from this program, which provides them with a better understanding of how seeds may respond to a site’s unique soil temperature and moisture regimes.

You can access the Auto-Germ Program, and the case study research by following these links. 

Dr. Matt Madsen is an assistant professor of Plant and Wildlife Sciences at Brigham Young University.
Dr. Bruce Roundy is a professor and range ecologist, retired from the plant and wildlife science department at Brigham Young University.
William Richardson is a graduate student, currently at University of Nevada, Reno.


Lisa Ellsworth
Project Co-coordinator
Dept. Fisheries & Wildlife
Oregon State University
Corvallis, OR  97331
(541) 737-0008

Beth Newingham
Project Co-coordinator
GB Rangelands Research
USDA Ag. Res. Service
Reno, NV  89512
(775) 784-6057 ext. 233

© Sagebrush Steppe Treatment Evaluation Project (SageSTEP)
Site Designed by Kite Media