Risk, Rate, and Impact of Medusahead Invasion of California Savannas

Final Report for GW07-006

Project Type: Graduate Student
Funds awarded in 2007: $19,971.00
Projected End Date: 12/31/2008
Grant Recipient: UC Davis
Region: Western
State: California
Graduate Student:
Principal Investigator:
Expand All

Project Information


This graduate student project is integrated into a pre-existing USDA Western SARE project (SW06-038) intended to develop grazing-based methods for medusahead (Taeniatherum caput-medusae) control. Using Landsat imagery and ground surveys across several counties, Mh coverage was mapped over 12 Mha of rangeland in the California Central Valley and surrounding foothills. Two small-scale studies indicated that Mh thatch accumulation has little effect on Mh invasion or removal, and that Mh is not competitive against ryegrass in the absence of defoliation. Spatial mapping of “precision grazing” and Mh distribution over treated pastures in Yolo County was conducted and available results are discussed.


Medusahead (Taeniatherum caput-medusae) is an invasive noxious grass that is unpalatable to livestock and wildlife and slow to decompose (Young 1992, George 1992). Accumulation of medusahead (Mh) residue forms a heavy thatch layer that may inhibit other plant species (Knapp 1998). By 1999, over 5 million acres in central and northern California contained Mh (Miller et al. 1999). Although it spread southward into 10 more California counties during the last decade, there is little original information about impacts or management of Mh. Only one study of environmental factors related to Mh invasion was published (Dahl and Tisdale 1975). As part of a larger parent project I initiated several studies to better understand the distribution of Mh and the processes related to its invasion and control.

Project Objectives:
  1. 1. Create a method to determine the risk that a pasture or a land management unit will be invaded by Mh. Use method to prioritize areas for prevention of new Mh infestation in participating ranches.
    2. Improve effectiveness of precision grazing to control Mh by determining how patch size and density affect utilization by livestock and subsequent Mh spread.
    3. Establish thresholds for Mh control by assessing the effects of Mh spatial distribution and patch size on loss of grazing capacity and biodiversity.


Click linked name(s) to expand/collapse or show everyone's info
  • Emilio Laca


Materials and methods:

To achieve Objective 1, I used a ground-based survey in the Dunnigan Hills region of Yolo County, California, to calibrate Mh level in Landsat satellite images from May and June 2008. Roughly 2000 observations were made of rangeland and other land cover types (wheat, safflower, vineyards, etc.) on a 60 x 60 m support, and the percent absolute coverage of Mh in each rangeland observation was estimated. Two quadratic discriminant analyses were fit to the observations to divide them into land cover class (rangeland or other) and to divide classified rangeland observations into classes of Mh coverage (<5%, 5-40%, and >40%). Because of differences in crop life-cycle and canopy density, rangeland was easily identified in independent hold-out observations (both in the Dunnigan Hills and elsewhere in the California Central Valley). Therefore, the remainder of this approach focused on selecting an accurate discriminant model for identification of Mh-infested rangelands and identifying predictions associated with relatively high levels of confidence in independent data gathered in regions outside the Dunnigan Hills (264 observations in Yolo, Solano, Butte, Tehama, and Tuolumne counties).

To achieve Objective 2, data were collected from precision grazing trials in Yolo and Glenn Counties in 2007 and 2008. Precision grazing trials were designed to reduce Mh seed productivity and coverage using intensive grazing targeted at the early reproductive phase of Mh. These trials consisted of 8 to 12 paddocks on scales from 0.2 to 1.5 ha and grazed by sheep or cattle to roughly 25 to 65% dry forage utilization levels over 10 to 20 days while Mh is in early reproductive stages. In each case, surveys of Mh coverage and seed productivity were or will be conducted at selected times before and after grazing treatments in each plot during the year of and year after treatment. Generally, surveys consisted of 30 to 70 observations (including Mh coverage, Mh seed number, and/or canopy height) on a 0.1 m2 quadrat in each plot, distributed along uniform grids.

To achieve Objective 3, two small-scale experiments were also conducted to test whether mechanisms related to Mh patchiness (thatch accumulation, density-dependent effects on inter-specific competition under environmental conditions favoring Mh) do in fact drive Mh invasion by increasing relative Mh fitness. Firstly, a small-plot experiment at an Mh-infested ranch was designed to test the effects of Mh thatch on coverage and seed productivity of Mh, non-Mh grasses, and forbs. In two infested sites (one open and one shaded), Mh thatch removal (yes or no) was factorialized with non-Mh seed addition (a mixture of annual and perennial species at 400% of the recommended rate; yes or no). At two non-infested sites (drained and riparian), Mh thatch addition (at a density of ~5000 or 0 kg ha-1) was factorialized with Mh seed addition (at a density of ~1500 or 0 seeds m-2). All treatments were replicated four times at each site. Secondly, a pot-study was conducted under controlled conditions factorializing different densities of Mh and ryegrass seeds (0, 500, 3000, and 10,000 seeds m-2) and environment (shallow soil with water availability skewed toward the beginning of the growing season, shallow soil with water availability skewed toward the end of the growing season, and deep soil with water availability skewed toward the end of the growing season). Treatments were replicated 2-4 times. Plant number, phenological development (leaf and tiller number and reproductive stage), inflorescence, and seed density and soil moisture content were monitored on a bi-weekly basis.

Research results and discussion:

Rangelands occupied roughly 3.1 of the 12.5 Mha of land in the three Landsat tiles in which observations were collected, mainly along the margins of the Central Valley and surrounding foothills. Only a model based on green, red, and near-infrared reflectances of Dunnigan Hills observations in May and June appeared to provide an unbiased model of Mh coverage class in independent observations in regions described above. High (>40%) Mh coverage could be assigned to little (1.7% or 0.6%) of the 3.1 Mha of rangeland area when prediction error rate was held to 15% or 5% (respectively). A greater proportion of rangeland was infested with high Mh coverage in the northernmost Landsat scene, and 10-20 “epicenters” of Mh invasion on scales of 0.005-0.05 Mha were distributed along the length of the mapped portions of the Central Valley. These results indicate Mh may continue to intensify in the southern portion of the valley, and effective control may require regional treatment plans corresponding to the Mh invasion epicenter. Maps of predicted Mh from these images will be used to model Mh level and changes in Mh level over time as functions of environmental factors (proximity to highways, soil type, topography, weather, and management).

Data analysis from the 2007 Yolo County precision grazing trial showed that stocking densities (4-14 AU ha-1) and durations of grazing (7-17 days) resulted in utilization levels sufficient to reduce mean canopy height well below that needed for drastic reduction of Mh seed productivity (canopy height in 89% of 1800 survey observations below 9 cm; a reduction of 75-80% in the year of treatment and in the subsequent year averaged across treated plots). Variation of apparent utilization levels and/or grazing duration within the ranges we studied had no significant effect on Mh control. In this situation, the impact of precision grazing on Mh was quite high, such that any spatial variation in forage utilization in response to Mh patch size and density was relatively unimportant. In contrast, higher forage growth than expected (following late precipitation) resulted in understocking sheep and cattle in experimental plots in 2008, such that livestock easily avoided grazing Mh regardless of patch size or density. Spatial analyses of these plots will require further data collection in 2009 (part of the larger parent project).

In my small-plot tests of thatch impacts, neither Mh thatch removal from invaded sites nor Mh thatch addition to non-invaded sites had strong effects on subsequent seed productivity of Mh, non-Mh grasses or forbs or on mean Mh seed mass, inflorescence density, or mean seeds inflorescence-1. In comparison, site effects and apparent alleviation of seed limitation exhibited much stronger effects on seed productivity of both Mh and non-Mh species as well as mean Mh seed mass. In Mh-invaded sites, thatch removal doubled the absolute coverage of forbs and (in the absence of non-Mh seed addition) quadrupled non-Mh grasses, but also increased the absolute coverage of Mh by roughly 50%. Although thatch effects on seed productivities were not strong enough to determine Mh invasion or control, thatch removal might still be valuable for improving forage availability because it more than doubled the absolute coverage of all living plants at both Mh-infested sites I studied (from 20.1 to 46.0% at the open site and 18.3 to 38.9% at the shaded site).

Preliminary data analysis from the pot study suggests that, in the absence of defoliation, Mh leaf coverage and tillering rate at any of the densities studied is highly reduced when ryegrass plant densities are roughly 100 individuals m-2 or more. In contrast, ryegrass leaf coverage, tillering rate and reproductive rate are reduced by Mh only when ryegrass individual density is very low (less than 50 individuals m-2) and Mh density is very high (1500 individuals m-2 or more). Per individual, ryegrass also reduced soil moisture content much more than Mh (roughly 2.5 times as much during early growth), resulting in dramaticaly less late-season soil moisture content where Mh was grown with ryegrass compared to Mh grown alone. Medusahead reproductive rates across the entire experiment were very low, perhaps due to insufficient vernalization following a late seeding date (early March). However, these results indicate that monotypic patches of Mh may play a critical role in invasion by effectively conserving soil moisture until Mh reproduction occurs.

Participation Summary

Research Outcomes

No research outcomes

Education and Outreach

Participation Summary:

Education and outreach methods and analyses:

Outreach efforts are integrated with the parent Western SARE project and include field days and presentations to weed management and growers’ groups. Publications currently drafted but not yet submitted to peer-review journals include manuscripts regarding the mapping of Mh across the Central Valley, the small-scale experimental test of thatch effects on Mh invasion and control, and the evaluation of precision grazing treatments at the Yolo County site in 2007. These manuscripts also form my PhD dissertation.

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.