Developing Sustainable Management Procedures for Widespread Noxious Weeds on Public Lands in the Colorado Front Range

Timothy Seastedt, Ph.D.
University of Colorado
Institute of Arctic and Alpine Research and Department of Environmental, Population and Organismic Biology
UCB 450
Boulder, CO 80309-0450
303-492-3302
303-492-6388
gary.brewer@ndsu.nodak.edu

Executive Summary

Invasive species of weeds continue to be a national concern. Conversion of rangelands and natural areas to ecosystems dominated by new species is both a major economic concern and an ecological crisis. In response to this problem, weed managers are being forced to deal with ever increasing weed populations. Very few weed managers feel comfortable using non-chemical controls since so little research has been done showing that these methods are cost-effective and that they work. Therefore, herbicides are often the most common tool being used to control widespread noxious weeds.

In Boulder County, we have a significant diffuse knapweed infestation on our public lands. One important and underutilized tool for non-chemical control of knapweed is biocontrol insects. We have successfully established very high densities of five species of biocontrol insects at a county-owned study site. Our preliminary results indicate that significant reductions in knapweed populations have been achieved on our ungrazed prairie area. However, the effects of many components of public lands management practices on these biocontrol insects remains unknown. As we now have an abundant supply of these insects, we propose to investigate effects of previous herbicide use, grazing by cattle and prairie dogs, fire, and finally, how different plant communities respond to both knapweed and if biocontrol impacts are a function of plant community type.

We believe that the use of biocontrol insects remains limited in part due to a concern by conservationists of nontarget effects of these insects. We have already initiated studies for the knapweed insects, and we propose to expand this activity to involve other insects now available in Colorado for weeds that demand management actions. We also propose to initiate studies of weed population dynamics in various plant communities to help establish exactly when and where these species need proactive management actions.

In our ongoing demonstration project and the project proposed here, our goal is to develop a weed management program that incorporates traditional integrated vegetation management methods with knowledge derived from ecosystem science. This approach attempts to manipulate the biotic and abiotic components of the environment to resist invasion of non-native weedy species. At the same time, our goal is to integrate the various weed management options with those management activities that favor the growth and survivorship of native species. An inevitable outcome of this process will be to integrate weed management within the framework of adaptive ecosystem management that will reduce the need for pesticide use on these public lands.

Objectives

To greatly reduce or eliminate the use of herbicides to control established, widely spread noxious weeds on public lands by use of biocontrol insects and other nonchemical management practices.
To develop a program that recognizes "weed population control" from a classical ecological perspective involves developing sustainable mechanisms that impose density-dependent population regulation on the target weeds. In contrast, programs that emphasize weed kill lack this critical control element required for sustainability.
To develop nonchemical weed control strategies that are easy to implement with the existing budgets, person-hours and equipment currently owned by weed management districts.
To determine how land management and weed management strategies affect the efficacy of biocontrol insects.
To conduct a rigorous analysis of the threat of selected widely dispersed weed species to specific plant communities.
To conduct extensive surveys and tests of potential biocontrol insect effects on nontarget species.

Justification

Management of widespread noxious weeds such as diffuse knapweed, Canada thistle, Dalmation toadflax and others requires procedures conceptually different from those used to manage new invading species. These widespread weeds have established, and these species are very likely to remain a part of our environment. While such species can be eradicated locally for brief periods, their reestablishment at some level is almost inevitable. However, it is not too late to development management techniques to minimize the negative impacts of these species without the use of extensive, recurrent use of herbicides. Here, we propose a partnership among four public agencies to facilitate a potential reduction in pesticide use by proactive testing of nonchemical management activities for weed control. The University of Colorado, The City of Boulder Open Space and Mountain Parks, and the Colorado Department of Agriculture will work together on a project supported by PESP funds from Region VIII of EPA. The goal of this project is to expedite the collection of information that will contribute to sound land management while reducing reliance on herbicides to control widespread noxious weeds on public lands.

Facts about wide-spread noxious weeds that argue for development of nonchemical controls:

Reintroductions are the rule rather than the exception. Containment using chemicals requires repeated applications. While spot spraying or wicking might be acceptable, large-area spraying is deemed an extreme, last resort to be used only if other methods are unavailable.
Noxious weeds will not pose threats to native species in parts of their geographic range and within certain plant community types, even though they will establish in such areas. Almost all land managers have observed plant communities where noxious weeds remain at most subdominants within the community. Herbicide treatment to such areas is potentially unwise and unnecessary. Identifying these habitats as ecosystems less vulnerable to invasion and harm by selected widespread noxious species is important.
In some but not all cases, biocontrol insects and pathogens offer the greatest potential for low-cost, sustainable control. Identifying these, and identifying legitimate problems (effects on nontarget species) is critical to science-based management decisions regarding their use.

Weed managers in Boulder County rely on the use of herbicides such as picloram to manage exotic weeds on city and county lands. In addition to roadsides, about 2000 acres of impacted native areas have been sprayed annually in recent years, at a cost to local taxpayers of more than $50,000 per year. While significant advances in the use of biocontrols for sustainable management of diffuse knapweed have occurred (Seastedt et al., in review), county and federal agencies such as DOE Rocky Flats continue to aerial spray large areas for knapweed. This suggests that a significant time lag exists between discovery and implementation of nonchemical methods. One way to speed this process is to actively involve personnel from the land management agencies in the nonchemical demonstration projects, as is proposed here.

Weed infestation threatens native plant diversity, but the spatial extent of this threat to native flora remains poorly documented. Meanwhile, herbicides can cause damage to native plants and can present a health threat to other organisms, including humans. Open Space and Mountain Parks lands have extensive visitor use, and the worst weed infestations tend to be in disturbed areas where people and pets are likely to come into contact with herbicides. Biocontrols come with their own set of risks to nontarget species (Strong and Pemberton 2000). Responsible use of these methods therefore requires that nontarget effects be known and deemed acceptable relative to risks of alternative methods and problems caused by the invasive weeds (Pimentel 2000).

Research at CU has taken an ecosystem science approach to the analysis of invasive species. (Seastedt et al. 1996, Reever-Morghan and Seastedt 1999, Reever-Morghan et al. 2000, LeJeune and Seastedt 2001, Byers et al., in press, Seastedt et al. submitted). This work has benefited from an association with the Long-Term Ecological Research (LTER) network, a program that for over 20 years has been involved in studies that assess plant species composition and abundance (e.g., Knapp and Seastedt 1998). These studies have been useful in identifying those characteristics that lead to invasibility of grasslands (e.g., Smith and Knapp 1999). By understanding those mechanisms that contribute to the dominance of native species, we gain insights into causal mechanisms for persistence and dominance, and what may be required for nonindigenous species to assume and maintain dominance in these communities (e.g., Davis et al. 2000; LeJeune and Seastedt 2001).

To understand the mechanisms that allow a species to become dominant within a community, one must understand characteristics of both the species and the ecosystem. Dominance is not a species trait; dominance is the result of species traits interacting with the resources and restraints provided by the host ecosystem. Weed science therefore needs a proper mix of population and ecosystem science. From a population ecology perspective, "control" of wide-spread weeds (those with well established seed banks) can be equated with mechanisms designed to keep a species from obtaining dominance or being maintained at levels where negative attributes of the species are not a serious concern. Such populations would be well below the maximum carrying capacity generated by intraspecific competition mechanisms (i.e., at densities where the invasive species becomes a monoculture, excluding all others). In a classical sense, density-dependent population mechanisms provide the means whereby species are maintained at some defined level by increasing mortality factors when densities are above some threshold level.

Programs that emphasize weed kill generally operate as density-independent mortality functions. The weed population grows exponentially until locally extirpated by herbicides or other mechanisms. Such mechanisms may maintain a weed population below some threshold level, but the exponential growth curves of weeds being managed in this manner actually argues that these populations are not controlled, just contained. In the case of herbicides, this form of containment is anticipated to eventually fail due to the development of herbicide resistance by the plant. This does not constitute long-term sustainable control. In contrast, density-dependent controls resulting from interspecific competition, predation, parasitism and pathogens are usually the products of natural selection. Biological controls operate most strongly when densities are high, thereby reducing but not eradicating the target plant. There is an important lesson to be emphasized here to weed managers. Under this system, the presence of the weed in low numbers somewhere on the landscape is essential to its management and sustainable control. An example for the Colorado Front Range is St. John's wort, (Hypericum perforatum), which has been controlled by the presence of the beetle, Chrysolina quadrigemina (W.A. Webber, personal communication).

A large body of research has demonstrated that plants are capable of affecting soil nutrient availability (e.g., Wedin and Tilman 1990) but that management of these plants using fire and grazing can also affect resources (e.g., Seastedt et al. 1991). By altering soil resource availability with biotic or abiotic methods or by affecting light availability by selecting for specific competing species, we can develop resource-limitations ("bottom-up controls") for invasive species (Figure 1). By enhancing rates of herbivory, seed predation, or pathogen abundance, we can impose "top-down controls" on these same species (Figure 1). These controls can be used singly or in combination. Our current knapweed work suggests that the top-down controls are effective at our primary research site, but we do not as yet know if this approach will be successful where resource availability differs substantially from our current study area.

Figure 1. Schematic of how an invasive species like knapweed can be controlled. Arrows leading to the knapweed (bottom-up controls) suggest that resources to the plant are minimized by enhanced competition from native species or from direct manipulation of soil resources. Arrows leading from knapweed (top-down controls) have been greatly increased by release of biocontrol insects.

Literature Review

A. Importance of the Problem

The spread of invasive species has become a serious problem in almost all regions of the country (Mack et al. 2000). There is sufficient concern about the spread of non-native species that the federal government created an interagency committee, FICMNEW (The Federal Interagency Committee for Management of Noxious and Exotic Weeds) to work cooperatively towards the prevention, control and management of exotic weeds, while maintaining and restoring ecosystems and biological diversity. The extent of the concern of the federal government about the problem of non-native plants was further underscored by an executive order from the White House in February of 1999 "to prevent the introduction of invasive species and provide for their control and to minimize the economic, ecological, and human health impacts that invasive species cause."

In many cases, pesticide application is the only method being used to address the problem of exotic weeds. Since repeated applications are necessary, the use of herbicides is not a permanent solution and presents risks to human health and to the environment. Although there is a movement towards Integrated Pest Management (IPM), weed managers often find themselves relying almost exclusively on pesticides due to a lack time, money, and adequate research that demonstrates that non-chemical methods are cost-effective, labor effective, easily understood and most importantly, work. Also, weed managers are seldom trained to work with insects, and while the learning curve for using insects is fairly simple and straightforward, there seems to be a moderate to strong reluctance towards using this potentially effective method. This may be due to concerns expressed by ecologists that non-target impacts of these insects may be significant (e.g., Strong and Pemberton 2000). David Pimentel (Pimentel 2000) and others believe this concern is overstated and lacks generality. However, there remains risk, however small, and there are simple procedures to test for nontarget effects by these biocontrols. Once these tests have been conducted, a much better understanding of risk will be established, and weed managers will be able to make science-informed decisions.

Diffuse knapweed (Centaurea diffusa) remains a significant pest in much of the western U. S. and is a serious problem in Boulder County. Much of the pasture land and prairies to the east of the Front Range has high densities of knapweed. Diffuse knapweed and other species in Centaurea are European in origin and were first introduced into North America accidentally in the 1900's (Muller-Scharer and Schroeder, 1993). Recent estimates for infestation of knapweed in Canada and the U.S. is 1,500,000 hectares with alarming rates of increase (Piper and Rosenthal, 1992; DiTomaso 2000).

Several characteristics of diffuse knapweed make it difficult to manage. Diffuse knapweed is usually a biennial, producing a rosette the first year. The second summer it bolts and flowers, although under crowded conditions, it has been reported that the vegetative rosette stage can persist for years (Powell, 1990). It produces a deep taproot that can reach deep water sources out of reach of competing grasses in dry conditions. Knapweed produces large numbers of seed, and once the plant is mature, it can become a tumble weed and travel great distances while dispersing seeds (Roche and Roche, 1988). The mature plant produces a sesquiterpene lactone called cnicin, which is bitter and toxic to non-adapted animals (Landau et al., 1994). Mature knapweed is unpalatable to grazing cattle (Strang et al, 1979). Even when knapweed plants have been killed and are not present, substantial seed banks give rise to new generations of knapweed. These seeds can survive for extended periods in the soil and are capable of continuous germination over the growing season (Sheley and Larson, 1996).

Knapweed is difficult to control even with herbicide treatment. Myers and Berube (1983) found that knapweed densities were reduced the season following herbicide treatment. However, a third of these treated plots contained knapweed seedlings and half had rosettes, suggesting reinvasion from the seed bank and/or herbicide resistance. We suspect that resistance to the herbicide of choice, picloram, which has been used for decades, is well underway if not already here.

Other wide-spread weeds that have attracted attention of weed and land managers include Dalmation toadflax, an escaped ornamental that now has several deliberately introduced and accidentally introduced biocontrol insect species available for its potential control in Colorado, Canada thistle, a major pest of more mesic areas in Open Space and Mountain Parks, and purple loosestrife, a major threat to wetlands. All of these species have biocontrol species that need to be tested against native, nontarget plants before they are likely to be well received by land managers. Leafy spurge is another widespread plant in Colorado that appears to be responding to the right species mix of biocontrol insects (Russell Johnson, Arapahoe County, personal communication). There are only very small patches of this species species on City of Boulder Open Space and Mountain Parks properties. Should the plant show signs of expanding on Open Space and Mountain Parks, we would include this plant and associated insects in our research plan. We will, however, work with Russell Johnson to make sure that his information is available to others.

B. Innovative control of knapweed and other weeds covering large areas

Learning to contain knapweed populations and reduce knapweed densities safely and effectively requires new methods and new ways of thinking. Although herbicides such as picloram can kill knapweed, reinfestation will occur unless other species are able to out-compete knapweed. This requires some kind of management technique to give the preferred plants a competitive advantage (Beck, 1994). Reseeding alone however, does not appear to always be an effective means of controlling knapweed (Larson and McInnis, 1989; Myers, 1996).

We have found that effective control of an established population of knapweed by biocontrol insects is possible (Figure 2; Seastedt et al. in review). We were able to reduce a field that was dominated by knapweed to levels where the relative plant cover by this species was less than 5% as of 2001. This level is below the threshold recommended for control by The Nature Conservancy (TNC 1998). However, we do not know a) if this control is sustainable, and b) how grazing by cows or prairie dogs or c) how prescribed burns affect this control. Measurements of these common management activities on the efficacy of biocontrol insects are therefore proposed for study here.

Unable to reproduce Figure 2 here.

Figure 2. A. Stem density, B. seedheads per plant, and C. seeds per seedhead observed on knapweed at a site where five species of biocontrol insects were released in 1997. The 2001 response occurred after insects increased to densities capable of having a strong negative impact on survivorship, growth, and reproduction of the knapweedEffective control of knapweed may be site-specific due to differences in climate, soil conditions and plant communities in different locales. This is why an understanding of how ecosystems affect the establishment and persistence of invading species is important in implementing IPM strategies. To our knowledge, no published studies exist that combine traditional IPM techniques with an ecosystem approach to weed management. We believe that this approach may be the only one that will offer long-term control of weedy species.

C. Background of the project

In April of 1997, the Boulder County Commissioners requested that a demonstration project take place on Open Space and Mountain Parks land for non-chemical control of diffuse knapweed. A 92 acre parcel of Open Space and Mountain Parks land was chosen, located just south of Superior in Boulder County. This site, the Lakota site, has been used for pasture land for decades and is in a degraded state from over-grazing. Cattle were removed at the time the study commenced in the spring of 1997 and no grazing has occurred since that time. During the summer of 1998, the project site area was expanded to include 160 total acres. This project has consisted of partnerships and community volunteer efforts. No funds were designated for the project. The university entered into a Memorandum of Understanding (MOU) with Boulder County. The County has provided in-kind support by using their equipment and personnel for mowing. Local nonprofits, such as the Sierra Club and Coloradoans for Alternatives to Toxics, as well as concerned citizens have donated their time for weed pulls and insect collection and release. The Colorado Department of Agriculture donated biocontrol insects for the project. University faculty, classes, as well as independent study students and graduate students have conducted studies and collected data.

In 1999 our group received EPA PESP funding to initiate a study designed to enhance the effectiveness of biocontrol insects. Results of this effort proved successful (Seastedt et al., unpublished report). Two important findings from that effort included 1) knapweed insects are attracted to extracts that include the knapweed chemical, cnicin, and 2) knapweed insects cannot be induced to attack other species even when these species are treated with this extract containing cnicin. We continue to explore possible uses and applications of this extract. Using cnicin extract to attract biocontrols to isolated patches of knapweed remains an untested management application of this finding. We plan to test this in early June of this year.

The removal of cattle helped in the recovery of native vegetation with the coinciding decline of knapweed. We also released several biocontrol insects at the site. In 1997, we released the root beetle (Sphenoptera jugoslavica), the seed moth (Metzneria paucipunctella), the seed weevil (Larinus minutus), and the root boring weevil (Cyphocleonus achates). Two tephritid flies, Urophora affinis and U. quadrifasciata are gall formers in the seeds of knapweed and arrived at the site on their own. All of these except for Metzneria have established in numbers where we're now collecting and distributing these to other sites. These insects have increased in numbers by many orders of magnitude (i.e.,200 Larinus weevils were released in 1997 and are now estimated in the tens of millions). These insects are now in the process of invading other knapweed-infested sites, including land owned and managed by the City of Boulder Open Space and Mountain Parks. It is this group that is now interested in continuing and expanding the efforts of using biocontrols. The city-owned "invasion site" is near the 1997 insect release site, and at least three of the five species (Urophora spp, and Larinus minutus) are known to be on the new site. We expect the root-feeders to arrive on their own this year if they have not already moved into the area. Documenting how fast these species can spread is significant. This new area offers burned and unburned patches of prairie, of areas treated with picloram in recent years versus untreated plots, and, along with existing County areas, contains areas with and without cattle and with and without prairie dogs. The opportunities to measure and understand factors that enhance or reduce biocontrol impacts on this important weed are large.

In summer 2001, Seastedt, LeJeune, and Suding at CU obtained a three-year grant from USDA to study plant-soil and plant competition components of diffuse knapweed. That study is well underway, and the group now has a much better understanding of nutrient use and competitive interactions of knapweed. That study does not, however, involve work with biocontrols, so there is minimal overlap with the proposed research. (Indeed, a problem has been the fact that the biocontrol insects have consumed some of the experimental manipulations!)

Rationale for biocontrol screening procedures on Boulder Open Space and Mountain Parks

The flora of Open Space and Mountain Parks and Mountain Parks is diverse. There are an estimated 1500 species of plants in Boulder County compared to 3000 for the entire state. Over 800 species occur in the 41,000 acres of OSMP—more than half the species for the entire county. This level of species richness for such a small area is especially noteworthy because portions of the county species are subalpine and alpine and not found on OSMP.

The flora of OSMP contains a number of relictual and endemic species and groups of species found in well-defined physical features such as outwash mesas, shale barrens, and foothills riparian areas. Individual species of interest also occur in unique communities such as mesic and xeric tallgrass meadows and eastern woodland communities.

Criteria for Native Species of Concern Related to Introduction of Biocontrols:

Close genetic relationship--select native members of the same plant family as the biocontrol host plant.
Endemic species of local distribution.
Rarity—agency, county, state, or federal rarity. Prevent increasing rarity of species and the costly implications of progressing to threatened or endangered status.
Dominance in ecosystem—examples—Penstemon virens, an endemic species, is one of the most common forest understory species. If toadflax biocontrols escaped to this plant, it could have important effects on community function. Likewise endemic, Aster porteri, is abundant on mesas and forests and decreases in cover would have implications for community composition.
Multiple groups of biocontrols such as knapweed and musk thistle—example—there are concerns that Rhinocyllus conicus is adversely affecting native Cirsium species. How might biocontrols for knapweed and Canada thistle compound this effect?
Particular value for wildlife—the native Cirsium species and a majority of other composites are important insect nectar and host plants. The loss or reduction of the Cirsiums and others like Chrysothamnus, rabbitbrush, or Eupatorium, Joe-pye weed, would have significant impacts on butterflies and other insects. Scrophularia lanceolata, figwort, and the Penstemon species likewise are important insect pollinated species that could be affected by toadflax biocontrols.

Additional Concerns:

Multiple insects per hosts, while increasing the efficacy of biocontrol, increase the possibility of impacts on native species singly or together. (we will certainly be able to test for the biocontrols of diffuse knapweed.)
Many of the native species of concern are so rare or infrequent, it may be difficult to produce enough plants for screening, especially in a short time-frame.
Local adaptation—if a native non-target species is only marginally capable of supporting biocontrol agents, could those biocontrols adapt over time in their ability to utilize the native? What are the long-term implications for non-target plants? If a low level of use of the native loosestrife, Lythrum alatum, is detected from the biocontrol insects used on purple loosestrife, Lythrum salicaria, could that use increase, especially as the non-native becomes highly colonized?
Altitudinal range—biocontrols may be targeted for plains ecosystems but may move into shrublands and forests, so that the full range of related native species should be evaluated.

Approach and Methods

Specific projects:

Diffuse knapweed: This study will 1) monitor the invasion of biocontrols onto City-owned lands, 2)document relative success of biocontrols across plant community types, with respect to grazing activities, and with respect to fire management, 3) monitor biocontrol response on sites colonized by prairie dogs, and 4) monitor relative insect abundance on areas previously treated with picloram versus untreated areas. Procedures will follow those developed by Seastedt et al. (submitted). This involves quadrat sampling of stem, seedhead, and seed abundance of the knapweed, and plant censuses for the presence and use of the specific biocontrols. We will replicate treatment plots wherever possible, and establish long-term monitoring areas in which the knapweed is destructively sampled. For individual treatment comparisons (fire - no fire, grazed - ungrazed, burned - unburned, herbicide - no herbicide) a replicated paired-plot approach (one that controls for site and site history effects) is deemed appropriate.
Briefly, stem and seedhead counts on knapweed are completed in the field. Seed abundance as well as abundance of seedhead insects are obtained by examining seedheads beneath a dissecting scope. Root-feeding insects are best censused on rosettes in late spring, when larvae are large and easily visible.

Knapweed may occur over the entire range of plant community types in Boulder County, from the most xeric grasslands to moist mountain meadows. Insects have now been released in all of these community types and their impacts and spread are under study here. Managers have noted that knapweed appears only marginally present in certain grassland types, including the xeric tallgrass prairie, and a xeric upland community rich in native forbs. We propose to quantify this pattern and assess if certain plant community types are relatively "non-invasible" with respect to diffuse knapweed. Such sites could then be removed from areas that have been subjected to the large-area herbicide programs.

An advantage of working on this city-owned property is that vegetation monitoring has been ongoing since 1996 (ESCO vegetation monitoring reports to City of Boulder Open Space and Mountain Parks). The City will maintain this monitoring effort and expand it, as necessary, to document knapweed abundance in various communities and document the vegetation response to the anticipated demise of the knapweed due to insect activity. Vegetation monitoring is done using a series of 50 m transects which record the presence of all species occurring along this transect, and calculates the absolute and relative abundance of species using a point intercept method.
Dalmation toadflax: Inventory existing release sites for use of native species by biocontrol insects. Monitor effect of biocontrol insects on target species.
Canada thistle: Monitor the abundance and impacts of Canada thistle in native habitats. Collect and quantify effects of current insect damage to seed production of this species.
Purple loosestrife: Monitor effect of introduced biocontrol insects on native plant species.

We propose to use existing biocontrol release sites to obtain as much of the necessary information on these species as possible. This simply involves observing the biocontrols and attempt to entice them to use nontarget plants at existing release sites. The proposed methods would involve temporary caging (or netting) of the insects on specific plants and observe what, if any use results from this forced interaction.

Organization Responsibilities:

City of Boulder Open Space and Mountain Parks

Provide long-term monitoring sites and conduct plant species inventories (species richness and species cover) to assess management activities with respect to noxious weed species.
Provide lists of "native plant species of concern" to be evaluated in terms of potential use by introduced biocontrols.
Participate in inspections of biocontrol release sites to assess for potential nontarget plant use by biocontrols. Assist in setting up tests of native species of concern at biocontrol rearing sites, and assist in providing data for a web site of Colorado biocontrol information. This activity will be supplemented by CU studies (see below).

Colorado Department of Agriculture

Identify release sites in the Front Range and nearby areas to be assessed for biocontrol abundance and biocontrol activity on target and non-target species.
Provide facilities to assist in feeding tests of biocontrols on native species.

University of Colorado

Provide inventories of weed densities and cover with respect to various management options.
Provide inventories of biocontrol abundance and activities.
Conduct studies of use of target and nontarget plants by biocontrols. Form partnerships with other agencies and groups (e.g., Rocky Mtn. National Park, CU herbarium, BCPOS, Jefferson Co., Arapahoe Co, CSU, etc). to make sure the list includes all species currently released in Colorado for widespread weeds (including those discussed above and leafy spurge), and that observations of biocontrols at Colorado release sites are made. CU personnel will be responsible for constructing a web site on Colorado biocontrols. This will be linked to the State of Colorado Dept Agri. Web site, which already contains some information on these insects. However, the goal of this site will be to provide comments upon the effectiveness of the insects and specific information about insects' use of nontarget plant. Initially, this web page will be created on Seastedt's web site (http://culter.colorado.edu/~tims/).

Seastedt is responsible for preparing all data and all reports associated with this project.

Impact Assessment

Assessment of biocontrols will be conducted in two ways. "Success" is judged by the reduction in stem densities and seed production of knapweed, and by the reduction in plant cover represented by the knapweed. The former data are gathered using quadrat counts and seed counts of the weed, while the latter is conducted using plant transect surveys (Seastedt et al., submitted). Another index of biocontrol activity is obtained by assessing insect use of the plants. This data is obtained as part of the seed count surveys, and from examination of rosettes.

Success of the program in terms of herbicide reduction can be assessed in terms of changes in large-area applications of herbicides on knapweed. We believe that a demonstration of biocontrol effects in 2002 and 2003 will greatly enhance the contention that sustainable control of diffuse knapweed by biocontrol insects is indeed possible. Prior to initiating this study, "…it appears that none of these (biocontrol insect) agents, alone or in combination, effectively controls diffuse knapweed populations." (TNC 1998). Our preliminary data (Figure 2) suggest this is not current information.

The criteria for a successful result for the second aspect of the study (investigations of Dalmation toadflax, Canada thistle, and purple loosestrife) are more open-ended. Ultimately, this will provide the information for management decisions regarding the use of specific biocontrols on City Open Space and Mountain Parks properties, should they be necessary. It is possible that significant problems with use of nontarget plant species by biocontrol insects may be discovered, at which time a legitimate "no use" decision would be anticipated. A better understanding of Canada thistle population dynamics on natural areas should provide the information when active management of any sort (herbicides or other methods) is necessary.

Finally, a successful outcome would be to see the merger of weed management activities within the context of adaptive ecosystem management. In other words, control for a specific species would be identified only within the context of management objectives developed to preserve or enhance preferred plant species. This is a blatant attempt to redefine weed management in natural areas as an activity whose primary function is to assist in promoting those plant species and communities we want rather than its current focus on simply removing species we do not want. To do this we must show that the management actions that select for knapweed reduction have at least neutral or positive effects on native species community composition. Those results can be garnered from measurements proposed here.

Literature Cited

Byers, J.E., S. Reichard, C.S. Smith, I.M. Parker, J.M. Randall, W.M. Lonsdale, I.A.E. Atkinson, T.R. Seastedt, E. Chornesky, D. Hayes, M. Williamson. In press. A call for research needed to reduce the impacts of nonindigenous invasive species. Conservation Biology.

Beck, G. 1994. Diffuse and spotted knapweed: Biology and management. Colorado State University Extension Service Report 3.110.

Davis, M.A., J.P. Grime and K. Thompson. 2000. Fluctuating resources in plant communities: a general theory of invasibility. Journal of Ecology 88: 528-534.

DiTomaso, J.M. 2000. Invasive weeds in rangelands: species, impacts and management. Weed Science 48: 255-265.

Knapp, A.K. and T.R. Seastedt. 1998. Grasslands, Konza Prairie, and long-term ecological research. Pages 3-15 in: A.K. Knapp, J.M. Briggs, D.C. Hartnett and S.C. Collins, eds. Grassland dynamics: Long-Term Ecological Research in Tallgrass Prairie. Oxford Univ. Press< New York.

Landau, I., H. Muller-Scharer and P.I. Ward. 1994. Influence of cnicin, a sesquiterpene lactone of Centaurea maculosa (Asteraceae), on specialist and generalist insect herbivores. J. Chem Ecol 20: 929-942.

Larson, L.L. and M.L. McInnis, 1988. Impact of grass seedings on establishment and density of diffuse knapweed and yellow starthistle. Northwest Science 63: 162-166.

LeJeune, K.D. and T.R. Seastedt. 2001. Centaurea species: The Forb that Won the West. Conservation Biology. 15: 1568-1574.

Mack, R.N., Simberloff, D., Lonsdale, W.M., Evans, H. Clout, M., and F. Bazzaz 2000 Biotic invasions: causes, epidemiology, global consequences and control. Issues in Ecology 5.

McCall, P.J., T.C.J. Turlings, J. Loughrin, A. T. Proveaux and J.H. Tumlinson. 1994. Herbivore-induced volatile emissions from cotton (Gossypium hirsutum L.) seedlings. Journal of Chemical Ecology 20: 3039-3050.

Muller-Scharer, H. and D. Schroeder, 1993. The biological control of Centaurea spp. in North America: Do insects solve the problem? Pesticide Science 37: 343-353.

Myers, J.H. 1996. Struggling with knapweed, a persistent, exotic invader. Antelope-Brush Ecosystem Symposium, CWS Technical report, Ottawa, CN.

Myers, J.H. and D.E. Burube. 1983. Diffuse knapweed invasion into rangeland in the dry interior of British Columbia. Canadian Journal of Plant Science 63: 775-783.

Papaj, D.R. and M. Aluja. 1993. Temporal dynamics of host-marking in the tropical tephritid fly, Anastrpha ludens. Physiological Entomology 18: 279-284.

Pimentel, D. 2000. Biological control of invading species. Science 289: 869.

Powell, R. D. 1990. The role of spatial pattern in the population biology of Centaurea diffusa. Journal of Ecology 78: 374-388.

Reever-Morghan, K.J. and T.R. Seastedt. 1999. Effects of soil nitrogen reduction on non-native plants in disturbed grasslands. Restoration Ecology 7: 51-55

Reever-Morghan, K.R., T.R. Seastedt and P.J. Sinton. 2000. Frequent fire enables ungrazed tallgrass prairie to resist invasion by Cirsium arvense (Canada thistle). Restoration Ecology 18: 194-195.

Roche, C.T. and B. F. Roche, Jr. 1988. Distribution and amount of four knapweed (Centaurea L) species in western Washington. Northwest Science 62: 242-253.

Seastedt, T. R., J. M. Briggs and D. J. Gibson. 1991. Controls of nitrogen limitation in tallgrass prairie. Oecologia 87: 72-79.

Seastedt, T.R., P. Duffy and J. Knight. 1996. Reverse fertilization experiment produces mixed results. Restoration and Management Notes 14:64.

Seastedt, T.R., Nathan Gregory and David Buckner. Submitted. Reduction of diffuse knapweed by biocontrol insects in a Colorado grassland. Weed Science.

Sheley, R.L., B. E. Olson and L. L. Larson. 1997. Emergence date effects on resource partitioning between diffuse knapweed seedlings. Journal of Range Management 50: 39-47.

Smith, M.D. and A.K. Knapp 1999 Exotic plant species in a C4-dominated grassland: invasibility, disturbance, and community structure. Oecologia 120:605-612.

Strang, R.M., K.M. Lindsay and R.S. Price. 1979. Knapweeds: British Columbia's Undesirable Aliens. Rangelands 1: 141-137.

Strong, D.R. and R.W. Pemberton. 2000. Science 288: 1969.

TNC (The Nature Conservancy). Element Stewardship Abstract for Centaurea diffusa Lamark, diffuse knapweed. (http://tncweeds.ucdavis.edu/esadocs/documents/centdif.html).

Wedin, D.A. and D.Tilman D. 1990. Species effects on nitrogen cycling: a test with perennial grasses. Oecologia 84: 433-441.

Timetable

August and September, 2002

Inventory weed and biocontrol densities on City of Boulder Open Space and Mountain Parks sites. Initiate studies of insect impacts on paired plots of
a) burned versus unburned prairie
b) grazed versus ungrazed prairie
c) prairie-dog colonized versus uncolonized prairie
d) sites treated with picloram in 2001-2002 versus sites not treated with herbicide (primarily to look at impact on biocontrol densities per plant, not to assess plant densities).
Travel to all relevant biocontrol release sites to evaluate potential impacts on selected nontarget species (see above list).
collect seeds of native plants for germination and growth in the greenhouse, to be tested against biocontrol insects in 2003 at the Palasades insectary site.

Autumn, 2002-Spring 2003

Summarize knapweed, insect, and plant community abundance data.

Summer, 2003

Repeat knapweed and biocontrol density inventories as in 1) above.

Fall of 2003 through Dec 31 (conclusion of project period)

Analyze data, summarize findings prepare final report and publications.

Major Participants

Project coordinator:

Timothy Seastedt, Professor
Department of Environmental, Population and Organismic Biology
University of Colorado
UCB Box 450
Boulder, CO 80309-0450

Major Cooperators:

City of Boulder Open Space and Mountain Parks
State of Colorado, Department of Agriculture, Plant Bicontrol Division

Project Budget

Budget Category	Grant Funding	Other Funding	Total Funding
Personnel
Project Coordinator:	4,642		4,642
Student Assistants:	17,460		17,460
Project Acctg. Asst:	1,188		1,188
Fringe Benefits	197		197
Project Coordinator:	1,048		1,048
Student Assistants:	115		115
Project Acctg. Asst:	258		258
Travel	1,800		1,800
Equipment
Supplies	200		200
Contractual
Other (Indirect Costs) On Campus: 47% of M.T.D.C.	12,789		12,789
Total	40,000		40,000