AN ABSTRACT OF THE THESIS OF
Alexander Park for the degree of Master of Science in Environmental Science presented on December 20, 2013.
Title: An Evaluation of Biological Control of Linaria dalmatica with Mecinus janthinus in Oregon
Abstract approved: ______Douglas E. Johnson
Dalmatian toadflax (Linaria dalmatica) has become a prolific invasive plant of rangelands in Oregon since its arrival in the early 20th century. In 2001, the Oregon
Department of Agriculture initiated a release program promoting the distribution of the stem-boring weevil, Mecinus janthinus as classical biological control agent to reduce densities of Dalmatian toadflax and improve ecological integrity. This retrospective study sought to answer [1] What impact has M. janthinus had on reducing densities of Dalmatian toadflax, [2] Is M. janthinus spreading naturally to new sites, [3] Should ODA still facilitate spread of the agent, [4] If the program is successful, what is the long-term benefit acquired from this biological control program? These questions were answered using historical biological release forms and monitoring data, and a 2013 state-wide survey of Dalmatian toadflax and M. janthinus. The results show that M. janthinus has reduced average Dalmatian toadflax densities from 9.45 ± 1.34/m2 prior to release, to 5.5 ± 1.1/m2 after release.
It was also found that the weevil has naturally migrated beyond their original release
sites, and that there was no correlation between distance from release site and abundance of the weevil. These results indicate that ODA and its partners no longer need to distribute M. janthinus. Predictive modeling indicates that the current infestation of Dalmatian toadflax has reached 1% of its mean biological potential, while further expansion will be limited by the presence of M. janthinus. Reduction of plant density through biocontrol is of particular benefit to cattle production in
Oregon where conventional control strategies of Dalmatian toadflax could reduce the annual value per acre of grazed land could be reduced by 6.5%.
©Copyright by Alexander Park
December 20, 2013
All Rights Reserved
An Evaluation of Biological Control of Linaria dalmatica with Mecinus janthinus in Oregon
by
Alexander Park
A THESIS
submitted to
Oregon State University
in partial fulfillment of
the requirements for the
degree of
Master of Science
Presented December 20, 2013
Commencement June 2014
Master of Science thesis of Alexander Park presented on December 20, 2013
APPROVED:
Major Professor, representing Environmental Science
Director of the Environmental Science Graduate School
Dean of the Graduate School
I understand that my thesis will become part of the permanent collection of Oregon State University libraries. My signature below authorizes release of my thesis to any reader upon request.
Alexander Park, Author
ACKNOWLEDGEMENTS
I would like to offer my sincere gratitude towards my supervisors at the Oregon Department of Agriculture Tim Butler and Tom Forney for making my Masters of Science possible, and the required flexibility that I have witnessed by them during this undertaking. Without their willingness to invest time and resources into their employees, expansions of the collective intellectual capital that has occurred under their leadership would not have precipitated as successfully.
To the ODA employees that helped collect this data over more than a decade – Bonnie Rasmussen, Glenn Miller, and Dan Sharratt – this wouldn’t have happened without your work. A sincere thank you to Gary Brown and Colin Park at USDA APHIS for helping to secure funding for this research and the work put in throughout the course of the M. janthinus biological control program.
To my mentors and committee members Eric Coombs and Dr. Doug Johnson, your wisdom and guidance represented the ball of thread that has led me out of the labyrinth. Eric, I would like to thank you for investing a substantial amount of your time and passion into this project. Your understanding of the complicated ecological systems at work in this project made it much more thorough, and humbled me throughout this process. Doug, your willingness to guide me through the complicated statistical, ecological, and administrative dilemmas throughout my time at this University allowed me to thrive. You are the quintessential Renaissance man, and someone who any aspiring scientist should try and emulate in their scientific and ethical endeavors.
Thank you to my fellow Masters Student Dan Esposito, who assisted me with some of the statistical analysis for this project. Cheers my friend!
To my Mom and Dad, thank you for supporting me through the ups and downs of my past endeavors, this one, and future ones to come.
This work was funded by the United States Department of Agriculture’s Animal and Plant Health Inspection Service; and the Oregon Department of Agriculture.
CONTRIBUTION OF AUTHORS
Eric Coombs from the Oregon Department of Agriculture designed and implemented the bulk the protocols and data collection for this project. He also assisted in querying and simplifying databases that constituted a large portion of this project. Dr. Doug Johnson from the Rangeland Ecology and Management Department facilitated in the design of the surveys conducted over the course of this project. He also assisted in the application of statistical theory of this research.
TABLE OF CONTENTS
Page
Chapter 1 – Introduction ...... 1
Objectives ...... 2
Chapter 2 - Literature Review ...... 3
Dalmatian toadflax ...... 3
History and Distribution ...... 6
Conventional Dalmatian Toadflax Management ...... 7
Biological Control of Dalmatian Toadflax ...... 8
Mecinus janthinus ...... 9
Predicting Modeling for Dalmatian Toadflax ...... 14
Chapter 3 - Materials and Methods ...... 16
Description of Study Area ...... 16
WeedMapper Dataset ...... 17
ODA M. janthinus Release Data ...... 17
ODA Biological Control Monitoring Forms ...... 19
Long-term Study Site ...... 20
Page
Survey Site Selection ...... 21
Potential Distribution of Dalmatian Toadflax Expansion in Oregon ...... 23
Data Analysis ...... 26
Uncertainty and Error ...... 28
KRESS Modeling Error ...... 29
Land Use Impacts ...... 30
Chapter 4 – Results ...... 32
History of M. janthinus Releases ...... 32
Terrain and Dalmatian Toadflax ...... 34
Release Data Analysis ...... 35
Vegetation Types and Dalmatian Toadflax ...... 36
Land Use and Dalmatian Toadflax ...... 38
Infestation Types and Dalmatian Toadflax ...... 40
Disturbance Regimes and Dalmatian Toadflax ...... 42
Weevil Dispersal and Abundance ...... 44
M. janthinus Site Occupancy ...... 46
Populations Trends ...... 47
Page
Dalmatian Toadflax Density Relative to Release Year ...... 50
Population Trends at Dufur Observation Site ...... 51
Ecological Amplitude of Dalmatian Toadflax ...... 58
Chapter 5 – Discussion ...... 61
Bibliography ...... 68
Appendices ...... 73
LIST OF FIGURES
Figure Page
2.1 Mature Dalmatian toadflax……………………………………………………………………………..…4
2.2 Healthy Dalmatian toadflax infestation…………………………………………………………..….6
2.3 Adult M. janthinus……………………………………………………………………………………….……10
2.4 Oviposition scars from adult M. janthinus females……………………………………………12
2.5 M. janthinus larvae boring……………………………………………………………………………..…13
2.6 Toadflax damaged by M. janthinus……………………………………………………………………14
3.1 Locations of Dalmatian toadflax locations in Oregon………………………………………...18
3.2 Basin wide watersheds……………………………………………………………………………..………23
4.1 History of the reported releases……………………………………………………………………….32
4.2 Releases of M. janthinus made per county……………………………………………………….33
4.3.a Terrain types…………………………………………………………………………………………………..34
4.3.b Estimated average density by terrain type……………………………………………………..35
4.4.a Vegetation types…………………………………………………………………………………………...37
4.4.b Estimated average density by vegetation type………………………………………………38
LIST OF FIGURES (Continued)
Figure Page
4.5.a Land use types……………………………………………………………………………………………….39
4.5.b Estimated average density by land use types…………………………………….…………..40
4.6.a Infestation types…………………………………………………………………………….……..……….41
4.6.b Estimated average density by infestation types…………………………….…………..…..42
4.7.a Disturbance types………………………………………………………………………………..………..43
4.7.b Estimated average density by disturbance types…………………….…………..…………44
4.8 Distance from release and abundance of M. janthinus………………………….…………..45
4.9 Weevil abundance relative to Dalmatian toadflax density……………….…………..…..47
4.10 Release density/m2 of selected weevil release sites over time………..……..……….48
4.11 Percent change from release density to 2013 density………………….………..……….49
4.12 Change in density/m2 of Dalmatian toadflax over time………….………………..….….50
4.13.a Density/m2 of Dalmatian toadflax and weevil abundance……….……………...……53
4.13.b Density/m2 of Dalmatian toadflax and percent plants infested…………...………..54
4.14.a 2003 Dufur site……………………………………………………………………………….……..…….55
LIST OF FIGURES (Continued)
Figure Page
4.14.b 2006 Dufur site………………………………………………………………………………….……….56
4.15 Biological wildfire……………………………………………………………….……………….………..57
4.16 Habitat suitability model for Dalmatian toadflax……………………………….….………58
An Evaluation of Biological Control of Linaria dalmatica with Mecinus janthinus in Oregon
Chapter 1 – Introduction
Dalmatian toadflax, Linaria dalmatica, is a competitive invasive species native to the
Mediterranean region. It escaped ornamental plantings in the United States in the early 19th century and invaded areas utilized for livestock forage, outcompeted native plant communities, and degraded wild lands (Lajeunesse 1999, Alex 1962). Because of its prolific seed production, aggressive vegetative expansion and habitat plasticity,
Dalmatian toadflax has become a major invader of rangelands in the West. In 2013, the Oregon Department of Agriculture (ODA) estimated Dalmatian toadflax impacted
352,955 gross acres in Oregon.
The toadflax stem weevil Mecinus janthinus Germar (Coleoptera: Curculionidae), was released as a biological control agent on Dalmatian toadflax in 2001 by the Oregon
Department of Agriculture (Coombs 2013). Adult M. janthinus feed on the leaves, in addition to larval boring of the interior of the stem. Release of M. janthinus has resulted in the reduction of Dalmatian toadflax densities in many areas across
Oregon. Twelve years of monitoring and this efficacy analysis were conducted to determine the efficacy of ODA’s biological control program utilizing M. janthinus to control Dalmatian toadflax.
2
Objectives
The purpose of this post hoc observational study was to determine whether M. janthinus has impacted Dalmatian toadflax populations across Oregon and the agent’s capacity to spread independently of prior releases. In order to address these issues, historical release and survey data collected by ODA entomologists and agronomists, United States Department of Agriculture Animal and Plant Health
Inspection Service (USDA APHIS), county, and local government services was analyzed. A 2013 statewide survey of release sites was used to compare against original release records to determine relative efficacy of M. janthinus. This observational study addressed four questions: [1] Has M. janthinus had any impact in reducing densities of Dalmatian toadflax in Oregon since inception of agent release;
[2] Is M. janthinus spreading naturally to new sites; [3] Should ODA still facilitate spread of the agent; and [4] If the program is successful, what are the long-term benefits acquired from this biological control program?
3
Chapter 2 - Literature Review
Dalmatian toadflax
Dalmatian toadflax, Linaria dalmatica (L.) Scrophulariaceae, is an erect, short-lived perennial herb that is native from southeastern Europe through southwestern Asia.
Cultivated for nearly four centuries in Europe for its attractive flowers, it subsequently escaped ornamental introductions as early as 1894 in the eastern
United States, and became a more widespread problem by the early 20th century in
North America (Vujnovic and Wein 1997, Alex 1962, Mack 2003).
In both the United States and Canada, Dalmatian toadflax has exhibited a strong proclivity for weediness that has resulted in it being listed as a noxious weed or weed seed in three Canadian provinces and in 12 U.S. states (Sing and Peterson 2011). This hardy, glabrous plant has a vigorous reproductive cycle both vegetatively and by seed
(Alex 1962). Germinating in the spring and fall, seedlings can rapidly establish 51 cm long taproot within eight weeks and produce two to five stems in the first season that can flower and set seed. In subsequent growing seasons they can reach up to 65 stems per plant (Figure 2.1.) (Wilson et al. 2005, Robocker 1974). Vegetative reproduction occurs from adventitious buds on primary and lateral roots that can produce independent plants which can extend 3.6 meters from the mother plant
(Alex 1962, Wilson et al. 2005). If pollinated, a mature plant can drop up to 500,000
4 seeds through the fall and winter, with seed a dormancy of 10 years (Vujnovic and
Wein 1997, Wilson et al. 2005, Alex 1962). Seed viability tests in a dry storage setting indicated a 76% germination rate of seeds stored from 1-5 years (Robocker 1970).
Figure 2.1. Mature Dalmatian toadflax absent any weevil damage. Photo Credit: Eric Coombs, ODA
Dalmatian toadflax most successfully invades areas of cultivation and/or soil disturbance where competition from other perennial plants is reduced (Figure 2.2).
5
Grasses do not appear to compete well with Dalmatian toadflax (Robocker 1974).
This success is attributed to its prolific creeping root system which effectively captures soil moisture (Robocker 1974). Because of its tendency to propagate in disturbed areas, it is generally affiliated with disturbed pasture and rangelands, roadsides, abandoned or unmanaged land, and gravel pits (Wilson et al. 2005). It also is associated with coarse-textured, gravelly soils in semi-arid regions where its effective capture of soil moisture generates the greatest competitive advantage
(Wilson et al. 2005, Robocker 1974). As an example of this competitive advantage,
Vujnovic and Wein (1997) noted that that where biological control suppressed St.
Johnswort (Hypericum perforatum L., Hypericaceae) in Washington and Idaho,
Dalmatian toadflax invaded the newly opened niche and also outcompeted native plants communities present at the site. Dalmatian toadflax’s ability to outcompete native vegetation impacts forage plants for livestock and reduces endemic plant species densities (Jeanneret and Schroeder 1992, Lajeunesse 1999). It is also somewhat toxic to livestock as it contains a glucoside antirrhinoside, a quinolone alkaloid, and peganine which cattle actively avoid consuming, displacing grazing cattle from areas infested with substantial forage (Lajeunesse 1999, Jeanneret and
Schroeder 1992). The sheer density of healthy, established Dalmatian toadflax populations can deter cattle from grazing infested areas as well (Lajeunesse 1999,
Jeanneret and Schroeder 1992).
6
Figure 2.2. Dalmatian toadflax infestation occupying niche space of a historically perennial grass community. Photo Credit: Eric Coombs, ODA
History and Distribution
The first records of Dalmatian toadflax in Oregon occur in 1908 in Multnomah
County, with a 1909 entry indicating its presence in the “Lewis and Clark Exposition
Grounds” in Portland, Oregon. The first spatially confirmable entry occurs in 1950 around the town Condon, Gilliam County, where a large population of the plant currently exists (Rice 2013). Currently approximated at impacting 350,000 gross acres, Dalmatian toadflax has proven to be one of the most expansive infestations by a noxious weed in the State of Oregon (ODA 2011). Because of its capability to
7 degrade agricultural and natural resources, control of Dalmatian toadflax throughout the State of Oregon began in 1983 when it was listed as a “B” rated noxious weed
(ODA 1983).
Conventional Dalmatian Toadflax Management
Conventional control strategies using herbicides and other removal techniques appear to have short-lived reductions in Dalmatian toadflax population densities.
Chemical treatments of Dalmatian toadflax require retreatment every 3 to up to 12 years in order to maintain low population densities, with recruitment occurring in the absence of active management (Lajeunesse et al. 1993). In addition, one of the most common and effective active ingredients for use on Dalmatian toadflax, picloram, can also damage native forbs and shrubs. Herbaceous plants provide more effective competition against Dalmatian toadflax than grasses (Robocker 1974). Removal of competitive herbaceous species during herbicide applications may not take into consideration the net benefit to ecological integrity associated with native herbaceous plants as competition to invasive species (Rinella et al. 2009). Use of picloram often leaves exclusive grass communities, which can become more prone to fire (Arnold and Santelmann 1966) which increase Dalmatian toadflax dominance at sites (Jacobs and Sheley 2003). In addition, the difficult country and terrain in which
Dalmatian toadflax has infested makes a successful state-wide management using
8 conventional control methods ultimately untenable. In conjunction, these multiple management issues made biological control of Dalmatian toadflax the most reasonable control option for state-wide control of this invasive species.
Biological Control of Dalmatian Toadflax
In 1983, the Oregon Department of Agriculture began releasing biological control agents to control Dalmatian toadflax populations across Oregon (Coombs 2013). The defoliating moth Calophasia lunula (H.) (Lepidoptera: Noctuidae) was released in
Deschutes County in 1983 and failed to establish. Since the failed releases, this moth has emigrated into Oregon from established populations in Washington and Idaho.
In 2001 both the root gall weevil Gymnetron linariae (P.) (Coleoptera: Curculionidae) and the stem boring weevil Mecinus janthinus were introduced. Gymnetron linariae was released into Wasco County, but was determined to have failed to establish by
2005. Mecinus janthinus became established in 2004 and began being collected and redistributed from several sites in 2007 (Coombs 2013).
The 2001 releases of M. janthinus were conducted in seven Oregon counties; in addition the weevil had already become established in Idaho by this time (Coombs
2013). In 2002, the weevil was recovered in two of the original seven counties,
Gilliam and Wallowa. Several of the release populations became collectible as
9 nursery sites in 2004, where they were redistributed to other infestations.
Reductions in stand densities of Dalmatian toadflax were noted by 2006 across
Oregon (Coombs 2013). Releases of M. janthinus continued at major infestations of
Dalmatian toadflax in 2007 by partnering with the Bureau of Land Management,
United States Forest Service, and county weed programs, because of emerging success witnessed in parts of Oregon. Collections and releases of M. janthinus continued in higher numbers through 2008 and 2009 throughout central and eastern
Oregon (Coombs 2013). In 2010, it was observed that M. janthinus was spreading independently of human releases in Harney County (P. Rasmussen, personal communication, November 3, 2013). In 2011, collection and redistribution by ODA of
M. janthinus was turned over to the county weed control programs to finish the implementation of the project (Coombs 2013).
Mecinus janthinus
The toadflax stem weevil Mecinus janthinus Germar (Coleoptera: Curculionidae),
(Figure 2.3) has become a successful classical biological control agent on Dalmatian toadflax in North America (De Clerck-Floate and Miller 2002). It is a univoltine stem- boring weevil specific to a limited number of Linaria species (Goulet et al. 2013).
Native to central and southern Europe, introductions of M. ianthinus for control of
Dalmatian toadflax began in 1991 in Canada (De Clerck-Floate and Miller 2002).
Between 1988 and 1990, P. Jeanneret and D. Schroeder conducted host specificity
10 tests at the International Institute of Biological Control in Delemont, Switzerland. M. janthinus was found to target non-native toadflaxes, specifically Dalmatian toadflax and yellow toadflax (Linaria vulgaris Miller) (Scrophulariaceae) which had become naturalized in North America (Jeanneret and Schroeder 1992). The biological control agent was found to have negligible impact on ornamental and native species tested
(Jeanneret and Schroeder 1992). There were 38 plants reviewed for host specificity as mandated by the Canadian Review Committee and the United States Technical
Advisory Group (TAG) on Biological Weed Control. Subsequent post-release research has shown that M. janthinus showed no evidence of non-target herbivory on native plants in greenhouse and field studies (Breiter and Seastedt 2007). M. janthinus was chosen because there were no native North American stem-miners that were natural enemies of introduced Linaria, and that they were expected to have more impacts than seed feeders and/or defoliators (Jeanneret and Schroeder 1992).
Figure 2.3. Adult M. janthinus on Dalmatian toadflax stalk. Photo Credit: Eric Coombs, ODA.
11
M. janthinus was found to be resilient in its adaptation to disparate environmental ranges. It was observed to occur just below the subalpine zones in the Alps to the
Mediterranean climate of Italy (Jeanneret and Schroeder 1992). In addition, the range of M. janthinus was predicted to occupy the same range as that of Dalmatian toadflax between the latitudes of 40° N and 52° N (Jeanneret and Schroeder 1992).
Two studies indicate that climactic factors limit population increases by M.janthinus.
De Clerck-Floate and Miller (2002) described how supercool events during winter minimums of between -28 and -35 °C in conjunction with the absence of an insulating snow layer to protect freeze intolerant adult M. janthinus resulted in 100% morality.
McClay and Hughes (2007) reporting suggests that insufficient warm days to allow for egg development through adulthood may also contribute to mortality of M. janthinus. Even with supercool events and subsequent high mortality at a latitude of
52° N, De Clerck-Floate and Miller (2002) found M. janthinus populations remained present on Dalmatian toadflax populations albeit in lower densities.
Dalmatian toadflax vigor is limited by M. janthinus. As an adult, M. janthinus feeds on the leaves and stems of the host plant in May to mid-July (Nowierski 2004). Eggs are deposited inside the stems of Dalmatian toadflax by adult females from June to mid-
July (Figure 2.4.). Jeanneret and Schroeder (1992) found that females lay an average of 1.15 eggs per day under laboratory conditions, but a more recent report by McClay and Hughes (2006) found that females laid 2.7 eggs per day under greenhouse
12 conditions. Larva develop between 23 and 34 days after oviposition, in shoots with a diameter greater than 0.9 mm (Figure 2.5.) (Nowierski 2004).
Figure 2.4. Oviposition scars from adult M. janthinus females. Red arrows indicate scars. Photo Credit: Eric Coombs, ODA.
Mining in the stems by the larvae has been found to cause premature wilting of shoots and suppression of flower development (Nowierski 2004, Goulet et al. 2013).
Adults emerge from overwintering in the stems in May. Goulet et al. (2013) found that when M. janthinus attacked Dalmatian toadflax, stems showed reduced apical growth, fewer fruits and flowers developed, and there was a lower biomass of stems compared to control plants (Figure 2.6.). The field research of the Goulet et al. (2013)
13 was conducted at a latitude of 47° N, and found that the attack of M. janthinus on
Dalmatian toadflax reduced both growth and plant reproductive potential (Figure
2.6.). A regional assessment of M. janthinus impact Dalmatian toadflax conducted by
Van Hezewijk et al. (2010) in British Columbia, Canada, found that increased weevil densities in stalks had a negative impact on their length. The weevil was found to be widespread throughout the range of Dalmatian toadflax in British Columbia regardless of intentional human releases, indicating a strong ability for natural spread after initial introduction.
Figure 2.5. M. janthinus larvae boring in the stalk of Dalmatian toadflax. Photo Credit: Gary Brown, USDA-APHIS
Emerging evidence from Toševski et al. (2013) suggests that the there may be presence of cryptic or sibling species within M. janthinus species complex. Mecinus janthiniformis is morphologically distinguishable from M. janthinus by only a few very
14 subtle characteristics. The redistribution of the M. janthinus complex may have inadvertently introduced both of these cryptic species, perhaps accounting for the successes and failures in efficacy of biological control projects on Linaria (Toševski et al. 2013). It was found that there was a prevalence of M. janthiniformis present on
Dalmatian toadflax, as compared to the propensity of M. janthinus to be present on yellow toadflax (Toševski et al. 2011).
Figure 2.6. Mature Dalmatian toadflax exhibiting damage from feeding by adult M. janthinus, and larval boring. Photo Credit: Amber Richman, USDA-APHIS
Predicting Modeling for Dalmatian Toadflax
In order to better understand the benefits of biological control as part of integrated pest management for its list of its noxious weeds, ODA began using habitat suitability models to predict potential ranges of weeds into the future, including Dalmatian toadflax. Understanding the potential range of specific weeds in the state allows for
15 a post-hoc comparison to the current infested areas that allow for analysis of efficacy of certain control strategies, including biological control. These models assist in determining where geographically a specific species can occupy an area based on environmental and topographic variables in which it has been identified (Johnson et al. 2005 a). Using predictive modeling, multiple factor analysis can be used to simultaneously take into account multiple factors that affect the potential habitable zone for Dalmatian toadflax (Johnson et al. 2005 b).
16
Chapter 3 - Materials and Methods
Description of Study Area
This study was conducted throughout Oregon where Dalmatian toadflax infestations were present, ranging from -123° to -116° W, and 41.9° to 45.9° N. The study area encompassed the breadth of the climactic zones Oregon contains, ranging from populations in temperate rainforests in Western Oregon, to sagebrush steppe in the
Northeast, and high desert in the southeast. Populations of Dalmatian toadflax occurred in areas with a minimum temperature of -15° C, maximum temperature of
44° C, maximum annual precipitation of 2040 mm, and a minimum annual precipitation of 208 mm. The lowest elevation Dalmatian toadflax population was located in the Willamette Valley near the Columbia River at 7 meters above sea level, and the highest was found near Strawberry Mountain in the Malheur National Forest at 2276 meters. However, this study was conducted east of the Cascade Mountain range that divides the State of Oregon, where much drier conditions paired with well- drained and coarse-textured soils, which create habitats where Dalmatian toadflax infestations flourish.
17
WeedMapper Dataset
The Oregon Department of Agriculture Plant Division, Noxious Weed Control
Program maintains a weed database that contains information on the locations of reported weed infestations throughout the state (Figure 3.1) (ODA 2011). This database includes weed locations from the United States Forest Service, Bureau of
Land Management, Soil and Watershed Conservation Districts, Cooperative Weed
Management Areas, County Weed Boards, and Oregon State University amongst other cooperators. Geospatial data that comprises the WeedMapper dataset for
Dalmatian toadflax began to be collected in 2000. The dataset has 4,891 entries from eight different managing agencies ranging from the local to federal level. This data can be accessed at www.weedmapper.oregon.gov/.
ODA M. janthinus Release Data
Beginning in 2001, releases of M. janthinus on Dalmatian toadflax populations were recorded using ODA’s biological release form (Appendix 1). There were 219 recorded releases reported since the inception of the program, detailing the number of weevils released in addition to data pertaining to Dalmatian toadflax, locational, and environmental data. Data from the form used in this analysis was: release year, county, latitude, longitude, number of agents released, terrain, vegetation type, land use, disturbance type, infestation type, gross acres impacted by invasive plant,
18 density per square meter of invasive plant. This data provides a unique wealth of information about infestations of Dalmatian toadflax, and their environment at the time of release of M. janthinus. We visited 11 sites release sites while, other regional
ODA staff provided the other nine site densities for the analysis in this study.
Figure 3.1. Locations of Dalmatian toadflax locations from the ODA WeedMapper dataset and release locations of M. janthinus in Oregon are displayed.
19
ODA Biological Control Monitoring Forms
ODA utilizes a standardized monitoring form when monitoring all of its biological control agents (Appendix 2). Data collected at each monitoring site included: date, weed, agent, county, latitude, longitude, manager, area, agent stage, agent abundance, invasive plant infested percentage, gross area infested, density of invasive plant, and notes. Many are self-explanatory, but some data collection types are specific to ODA’s rapid site assessment method. Agent abundance estimates for
M. janthinus populations was recorded in observable adults per minute of inspection search time. Part of the reason for using weevil abundance per minute was to determine if a site was able to collectible for later redistribution of the weevil.
Populations that exhibited 10 or more weevils per minute were deemed collectible for redistribution to other infestations. Infested percentage of invasive plant indicates an estimate of how many plants across the site were infested by M. janthinus or bore impacts from their presence such as oviposition scars or feeding damage. Gross area infested indicates an estimate of the observable area that is infested by Dalmatian toadflax. Density of invasive plant indicates the range of densities across the observable site, and an average across the site. These observational values are used as a baseline determination for estimating the decline and extent of each toadflax infestation.
20
Long-term Study Site
A study site was visited on a near-annual basis post-release of M. janthinus in order to record data on weevil and Dalmatian toadflax populations. This site was not set up as an experimental design prior to the release of the M. janthinus, but was used to quickly determine annual or near-annual baseline readings of both weevil and
Dalmatian toadflax populations. ODA used this methodology for data collection because of the limited personnel and fiscal resources that could be applied to any individual site, but preferred to use rapid site assessments to allow for broad analysis over a large area and time. This type of analysis was designed to detect change in target plant density by orders of magnitude over time (Coombs et al. 1996).
The site which had been analyzed for long-term population data occur near Dufur,
Oregon at 45.4002° N and -121.2694° W. The Dufur site had the original release occur in 2001, but the release was found to have failed because the host had been mowed by the landowner in 2002, thus an additional release of 500 M. janthinus was made in 2003. Data from this site were analyzed for trends in population abundance of weevils compared to the densities of Dalmatian toadflax. Observations were not recorded in 2005 and 2012.
21
Survey Site Selection
The large geographical area and relatively small timeframe for adult M. janthinus surveys in the late spring of 2013 necessitated a randomized linear feature survey (D.
Johnson, personal communication, May 10, 2013). Linear feature surveys utilize easily accessible thoroughfares such as roads and Off Highway Vehicles (OHV) trails in order to reach as many sites as possible across large areas to achieve the most amounts of data as possible within a limited time. A linear feature survey allowed for travel by four-wheel drive vehicles that increased the study’s efficiency. Major watersheds with large Dalmatian toadflax infestations were used as natural isolates within the geospatial distribution of Dalmatian toadflax populations (Figure 3.2.). The major watersheds used as part of this survey were part of the 221 subregion hydrologic units titled HUC 6, that include areas that are drained by a river system, groups of streams forming a coastal drainage area, a closed basin, or a reach of a river and its tributaries (Seaber et al. 1987). The linear features were randomized internally within these basin wide HUC 6 watersheds which included: Middle
Columbia, John Day, Deschutes, Oregon Closed Basin, Klamath, and Upper
Sacramento.
Timing of the survey was dictated by the phenology of both Dalmatian toadflax and the weevil. A 10-50% bloom by the toadflax was preferred to see it on landscape, coinciding with adult M. janthinus being active and while oviposition was in progress.
22
It was also important to survey while damage by the weevil on toadflax was evident.
The time that best fit these desired phonological characteristics was late June into late July. The 2013 survey was conducted between June 17th, 2013 and July, 17th
2013.
In each of these watersheds, linear features that were in close proximity to
WeedMapper or biological agent release data points were identified, digitized, and randomized for survey. These data points were then exported to spreadsheets that were used in conjunction with a GPS to locate sites in the field. In addition to
WeedMapper and former biological agent release sites visited, any Dalmatian toadflax populations found randomly along the linear feature being surveyed were also recorded on the monitoring form. If infestations were continuous amongst the random sites, populations were recorded every five kilometers for sake of efficiency.
If any roadside infestations were located outside of designated survey watersheds on way to another watershed, they were recorded as random sites, and recorded. In the course of three weeks of survey, 72 sites were visited, within all six watersheds.
23
Figure 3.2. Basin wide watersheds used as isolates in the randomized linear feature survey shown with the 2013 survey sites also identified. These watersheds are based on HUC 6 features.
Potential Distribution of Dalmatian Toadflax Expansion in Oregon
Working with Oregon State University beginning in 2011, ODA quantified the potential distribution of Dalmatian toadflax by utilizing topographic and climatic data in Oregon in a Multi-Criteria Decision Analysis (MCDA) Weighted Sum Model (WSM) inside the Kinetic Resource and Environmental Spatial System (KRESS) developed at
OSU (Johnson et al. 2005 a). Seven climatic variables and a digital elevation model
GTOPO30 were scaled to 256 levels and re-sampled to 1.25 arc-minute resolution, or
24 approximate 4 km2 for analysis in KRESS. Relative probabilities were extracted from where a condition is met, such as the presence of Dalmatian toadflax and used as the relative weight in the WSM (Johnson et al. 2005 b). Each variable was weighted equally and the potential risk of invasion was assigned in proportion to the distribution curve for that variable based on intersections between weed locations and environmental variables. Those areas most climatically and elevationally similar to current infestations were assumed to have the highest risk of infestation within the WSM (Johnson et al. 2005 a).
The environmental variables being utilized in the model consist of (1) GTOP030
Digital Elevation Model, (2) freeze free days, (3) growing degree days over 10 degrees
C, (4) precipitation, (5)average temperature maximums, (6) temperature means, (7) average temperature minimums, and (8) number of wet days. Climatic data was obtained from The Climate Source, Inc., whom developed the datasets using the
Parameter-elevation Regressions on Independent Slopes Model (PRISM). PRISM utilizes point measurements of environmental data, digital elevation models, and other geospatial data to generate annual and monthly climate data. The datasets utilized in KRESS were generated from data between 1971-2000 (Daly and Taylor
2001). Elevation data was obtained from the US Geological Service, Earth Resources
Observation and Science (EROS) Data Center and was comprised of the GTOPO30 data set (USGS 2012).
25
The modeling process for plant habitat suitability modeling consists of the following steps:
1. Define the area in which the plant currently exists through use of (Global
Positioning System) GPS, expert knowledge, and GIS
2. Identify the factors of importance (these being environmental and landscape
variables)
3. Build the GIS layers of factors that are needed as ASCII Raster Maps
4. Scale each of the factors between 0-256 so that they can be treated similarly
5. Determine or estimate of the “importance” or the weight of each of those
factors for mathematical analysis
6. Determine the spatial and temporal relationships between the factors
7. Build the model in the KRESS modeling interface
8. Process the weighted factors mathematically using a Weighted Sum
Algorithm
9. Each cell in the area being modeled is evaluated for “suitability” or conformity
with conditions at sites with known populations
10. View the spatial pattern of the model
11. Evaluate the model using statistical methods or in-field verification
The KRESS multiple factor analysis is used to simultaneously take into account a series of factors that affect the presence of plants for a particular position on the
26 landscape based on a deterministic application of rules (Johnson et al. 2005 a). A scientist or resource manager can conceptualize linear, non-linear, or mixed models, and if spatial data exists for the parameters chosen, apply them to the landscape. The user can then incorporate information about the system to build a model that seems reasonable and generate the suitability for each cell on the landscape (Johnson et al.
2005 a). The KRESS model was used in this research to quantify and convey the potential area for continued, unfettered expansion of Dalmatian toadflax populations given no biological control program was implemented by ODA.
Data Analysis
The 2013 survey weevil abundance and distance from release sites data was analyzed to determine whether M. janthinus was expanding its range independently of human facilitation. Distance from release was sites was determined using a straight-line distance calculation in ArcGIS, and was generated by using historical release record sites proximity to WeedMapper and random sites visited in the 2013 survey. It was hypothesized that M. janthinus had spread throughout the range of Dalmatian toadflax infestations in Oregon. If so, there should not be a correlation between distance from release sites and abundance of weevils. Alternatively, if there is a negative correlation; then distance from original release sites would be a limiting factor in weevil abundance. The abundance of weevils was log-transformed prior to
27 analysis to reduce impacts of outliers within the dataset. A linear model was developed to test the hypothesis against the alternative.
A linear model was also used to test whether there was a correlation between weevil abundance and density of Dalmatian toadflax. The hypothesis being that there is a correlation between weevil abundance and density of Dalmatian toadflax, with density being the limiting factor. Again the abundance data was log transformed prior to the running of the linear model.
Two analyses were used in order to assess whether M. janthinus had any impact on
Dalmatian toadflax densities in Oregon: (1) Densities of toadflax prior to release (n =
20 sites) were compared to densities after releases using a one tailed t-test, and (2) densities of toadflax relative to release year zero were compared against years from release using the Kruskal-Wallis non-parametric test, with inter-year determination of significance using the Bonferroni method.
Release form data analysis consisted of comparisons between different environmental and disturbance types releases were made on, as well as those categories analyzed against estimated Dalmatian toadflax densities using descriptive statistics. A Kruskal-Wallis non-parametric test was utilized to determine significance
28 between different categorical variables, and the Bonferroni method was used to determine inter-categorical significant differences.
Uncertainty and Error
Of the 72 sites visited during the 2013 survey, 54 sites were found to have Dalmatian toadflax and M. janthinus populations present. Eighteen sites that were designated for survey were either not found because Dalmatian toadflax became locally extinct and not detectable, observation error, GPS discrepancies or sites that had become extinct. These 18 sites were not included in analysis as their addition would previous erroneous and skewed results, and their biological relevance to the study was absent. Release and monitoring records were filled out by different observers with varying levels of ecological expertise, and thus significant human error could be contained within the data sets derived from these release records. Toadflax densities throughout this study were made as estimations of total site density in order to capture site trends, thus inaccuracies may have resulted from said monitoring method.
Release forms (Appendix 1) contained ranges of Dalmatian toadflax densities that that were designed to encourage persons conducting releases to easily indicate site densities (this was a response to low completion rate of filling out site densities on
29 release forms). When the person conducting releases did not write an integer down for an average, but instead indicated a range, the median value of that range was used as the average site density of Dalmatian toadflax. If no density was provided, we used 1 plant per square meter as a conservative minimal density at which a release would be conducted.
KRESS Modeling Error
The geospatial data collated to comprise the WeedMapper data points for Dalmatian toadflax was created from 11 different management agencies, with different collection protocols between and within the agencies. Disparate protocols create gross errors in both quality of observations, GPS precision, and quality control.
Additionally, each data point can represent one to and unknown quantity of
Dalmatian toadflax, neither indicating density of infestation or size. The datasets that were retrieved from the agencies also came in different geographic projections, which were subsequently transformed into WGS84 when conducting the standardization of the dataset, which will also generate spatial error. This is a best- case analysis for the present, until new methods with less error are developed and implemented.
Point data in the Dalmatian toadflax dataset did not consistently contain attribute indicating the size and density of the infestation, thus each presence was treated
30 equally where the whole cell was converted to a 1 to indicate presence. Each cell is approximately 394 hectares (973 acres), and thus the conversion to raster is a gross over-estimation of actual area infested by Dalmatian toadflax, but as each cell needed to be of exact size and dimension, this was necessary to implement the model and analysis.
The resampling and scaling process introduces error through bining in the case of scaling the environmental data, whereby a single value in the scaled dataset can represent a range of data from the original environmental dataset. Depending on the size of the continuous environmental dataset, the amount of values bined into a scaled value can vary. This makes analysis less precise as the scaled data represents one or more real-world values. The GTOPO30 data set was resampled to match the cell size of the PRISM dataset by averaging approximately 4 GTOPO30 cells to fit the
1.25 arc minute resolution of the PRISM data.
Land Use Impacts
A statewide land cover grid created by the Oregon Natural Heritage Information
Center was utilized in analyzing intersections between the habitat suitability model of Dalmatian toadflax (ONHIC 2010), and particular resources that are susceptible to invasion. These land use types were chosen by the ODA (ODA 2013), and by analyzing release forms in this study. The land cover grid was altered from its original
156 separate land use elements, and concatenated into resource categories based on
31 their vegetation type or land use. For this study, the appropriate elements were combined to display the general distribution of rangeland, right-of-way, and riparian zones. ArcGIS was used to overlay the mean of the habitat suitability model onto these particular resource categories to generate acreages of potential impact if
Dalmatian toadflax were to reach its mean ecological amplitude in these resource areas. It is important to note that because vegetation categories were used to generate these acreages, it does not reflect the political boundaries that define these lands utility i.e., areas considered rangeland with available forage may not be grazed by the land manager.
32
Chapter 4 – Results
History of M. janthinus Releases
Beginning in 2001 with seven releases of M. janthinus, the program expanded to the maximum annual number of releases reaching 57 in 2007, and subsequently trending downward to the minimum annual number of releases in of two in 2011 (Figure 4.1).
A total of 219 reported releases were made, distributing 71,119 total weevils in
Oregon between 2001 and 2012. Nineteen out of the 36 counties in Oregon had M. janthinus released within their boundaries; which targeted all of the known major infestations across the state. The year 2009 witnessed the largest number of weevils released with 25,950 weevils released across Oregon.
30000 60
25000 50
releaseed Weevils Released
20000 Number of 40 Releases
15000 30
10000 20 M.janthinus Releases
5000 10 Numner of individual Numnerofindividual janthinus M. 0 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Figure 4.1. History of the reported releases of M. janthinus by number of releases and number of individual weevils by year in the State of Oregon.
33
The largest number of releases, 46, occurred in Grant County; while the most number of weevils released in any one county was Harney County at a count of 16,750 (Figure
4.2). Harney County also had the highest average weevil count per release.
50
45
40
35
30
25
20
15
NumberofReleases Made 10
5
0
County of Release
Figure 4.2. Releases of M. janthinus made per county between 2001 and 2012.
34
Terrain and Dalmatian Toadflax
Foothills were the most common terrain type where M. janthinus releases were conducted, which represented 43% (n = 94) percent of all the releases (Figure 4.3.a).
The least common terrain type was Foothill – River with less than 1% (n = 1) of releases conducted in these terrain types. The terrain type with the highest average density per square meter of Dalmatian toadflax was mountainous terrain, with an average (n = 15) of 14 ± 2.75/m2 (Figure 4.3.b). There was significant difference between terrain types (Kruskal-Wallis, p-value > 0.01).
50%
40%
30%
20% Percantage oftotal releases 10%
0%
Figure 4.3.a. Percentage of different terrain types identified M. janthinus was released (n = 217).
35
16
14
12 2
m 10
8
6 Plant Plant density/
4
2
0
Figure 4.3.b. Estimated average density of Dalmatian toadflax/m2 by terrain type. Standard error bars shown in black.
Release Data Analysis
Data contained in the release forms generated from the inception of the program is illuminating in regards to the relationship between the density of Dalmatian toadflax infestations in relation to geo-physical, land use, and ecological variables. The most interesting relationships were between terrain, vegetation type, land use, infestation type, and disturbance types relative to the density of Dalmatian toadflax. This data illustrates environments at time of release of the biological control agents. Though no site is a statistical measurement by itself, when many are taken collectively over a large region, trends begin to become manifest.
36
Vegetation Types and Dalmatian Toadflax
Grasslands, shrublands, and a combination of both represented 65% of all vegetation types (n = 149) of which M. janthinus releases were made on Dalmatian toadflax
(Figure 4.4.a.). This vegetation type complex is representative of typical infestations of Dalmatian toadflax east of the Cascade mountain range. The vegetation type with the highest average density per square meter of Dalmatian toadflax was also grassland (n = 73), with an average of 9.3 ± 0.97/m2 (Figure 4.4.b.). The relationship between grasslands and the highest density of Dalmatian toadflax may be indicative of toadflax’s propensity to outcompete grasses more effectively than other vegetation types (Robocker 1974). There was significant difference between vegetation types (Kruskal-Wallis, p-value > 0.01).
37
40%
30%
20%
Percantage oftotal releases 10%
0%
Figure 4.4.a. Percentage of different vegetation types that M. janthinus was released upon (n = 218).
38
12
10
8 2
6 Plant Plant density/m 4
2
0
Figure 4.4.b. Estimated average density of Dalmatian toadflax/m2 by vegetation type. Standard error bars shown in black.
Land Use and Dalmatian Toadflax
The rangeland land use type (n = 73) independently held the propensity of all types which represented 54% of all the releases (Figure 4.5.a). Rangeland land use type generally correlates to vegetation types (grassland, shrubland) that Dalmatian toadflax has shown a strong propensity to compete against (Figure 4.4.a). The land use type with the highest average density per square meter of Dalmatian toadflax
39 was cropland with an average (n = 2) of 20/m2 (Figure 4.5.b.). Because cropland land use had only two observations of the same value, no standard error was available.
There was significant difference between land use types (Kruskal-Wallis, p-value >
0.01).
60%
50%
40%
30%
20% Percantage oftotal responses 10%
0%
Figure 4.5.a. Percentage of land use types where M. janthinus was released upon (n = 218).
40
25
20 2 15
10 Plant Plant density/m
5
0
Figure 4.5.b. Estimated average density of Dalmatian toadflax/m2 by land use type. Standard error bars shown in black.
Infestation Types and Dalmatian Toadflax
Patchy infestations (n = 123) were the most common infestation type where M. janthinus releases were conducted, which represented 57% (Figure 4.6.a). The least common infestation type was continuous linear with 1.8% (n = 4) of releases. The infestation type with the highest average density per square meter of Dalmatian toadflax were continuous infestations (n = 39), with an average of 8.23 ± 1.3
41 plants/m2 (Figure 4.6.b). There was significant difference between infestation types
(Kruskal-Wallis, p-value > 0.01).
70%
60%
50%
40%
30%
20% Percantage release total of
10%
0% Patchy Continuous Linear Isolated Patchy Linear Linear Continuous
Figure 4.6.a. Percentage of infestation types that M. janthinus was released upon (n = 216).
42
12
10
8 2
6 Plant Plant density/m 4
2
0
Figure 4.6.b. Estimated average density of Dalmatian toadflax/m2 by disturbance type. Standard error bars shown in black.
Disturbance Regimes and Dalmatian Toadflax
Grazing was the most common disturbance type (n = 80) where M. janthinus releases were made, which represented 36% of releases (Figure 4.7.a.). Disturbances that involved grazing of any kind (n = 120) represented 54% of the total all disturbance types. The least common disturbance type was logging operations with 2% (n = 6) where releases were conducted across the landscape. The disturbance type with the highest average density per square meter of Dalmatian toadflax were grazing and
43 road combinations (n = 10), with an average of 8.9 ± 2.8/m2 (Figure 4.7.b.). There was significant difference between infestation types (Kruskal-Wallis, p-value < 0.01).
40%
30%
20%
10% Percantage oftotal responses 0%
Figure 4.7.a. Percentage of disturbance types that M. janthinus was released upon (n = 217).
44
16
14
12
2 10
8
6 Plant Plant density/m
4
2
0
Figure 4.7.b. Estimated average density of Dalmatian toadflax/m2 disturbance type. Standard error bars shown in black.
Weevil Dispersal and Abundance
Of the 54 sites surveyed in 2013, 43 sites comprised of 23 random, and 11
WeedMapper sites were analyzed for a relationship between distance from release sites and abundance of M. janthinus. This excludes 20 release sites visited during the survey, because including release site abundance of weevils could skew correlation analysis by including both treatment and non-treatment. All sites that had Dalmatian toadflax also had varying levels of M. janthinus, with one exception; a right-of-way
45 zone which had been treated with herbicides and heavily disturbed. The most remote site recorded in the 2013 survey was approximately 60 kilometers away from the nearest known release, but the median distance was 1.5 kilometers from the nearest release. Using log transformed weevil abundance data, a linear model showed a weak negative correlation between abundance and distance from release site (Figure 4.8.) (r = -0.009, p-value = 0.41). Therefore the null hypothesis that there is no correlation between abundance of M. janthinus and distance from release site must be accepted.
M. janthinus Abundance and Distribution 450
400
350
300
250 Random 200 WeedMapper
countperminutesearch time 150
100
50 M.jathinus
0 0.1 1 10 100 Distance from release site (kilometers) Figure 4.8. Relationship between abundance of M. janthinus and distance from nearest release site. Release sites included in figure are non-zero because observations may have not been in the geographic center of the infestation.
46
M. janthinus Site Occupancy
The 2013 survey revealed that M. janthinus populations were present on Dalmatian toadflax infestations regardless of plant density. The average weevil abundance across all sites observed (n = 54) was 90 ± 12.2 weevils per minute of search time
(Figure 4.9). Though this data was recorded as a snapshot in time of a dynamic predator-prey relationship, the threshold of weevil abundance indicates that M. janthinus remains present on varying densities of Dalmatian toadflax populations.
These observations also indicate the strong reproduction potential of M. janthinus to respond to varying densities of its host plant across the landscape. Using log transformed weevil abundance data, a linear model was used to determine that there was a weak negative correlation between abundance and density of Dalmatian toadflax (r = -0.005, p-value = 0.41). It was found that the hypothesis of correlation between weevil abundance and Dalmatian toadflax density was not supported, and that there was no correlation between these two variables.
47
1000
100
10
1 Weevil Weevil abundanceperminute search time
0.1 0 2 4 6 8 10 12 14 16 Average Plant Density Figure 4.9. Weevil abundance relative to Dalmatian toadflax density data shown from the 2013 survey.
Populations Trends
95% of release sites monitored in the 2013 survey (n = 20) were found to have declining Dalmatian toadflax densities relative to the original release time density
(Figure 4.10). The mean reduction of Dalmatian toadflax across monitored release sites was 50% from time of release (Figure 4.11). The one site of increasing density was located in a right-of-way where human disturbances may be disrupting natural downward trends of Dalmatian toadflax impacted by M. janthinus. Reducing toadflax populations with mechanical or chemical controls reduces weevil numbers by
48 eliminating their food source and breeding location for one year, subsequently reducing their ability to control future infestations (Appendix 4).
30 Release Density 25 2013 Density
2 20 /m
15
Plant Plant Density 10
5
0
OIT
Klatt
Tumalo
1815Rd
Mcallister
CellTower
Hwy140 W
GaslineRdII
Skylinerpit 3
Harper Creek
ServiceCreek
HAGLESTEIN1
DevineCanyon
GoldenPasture
ParadiseHills/Bald…
Hwy140/coffee river WillowCreek (before…
Release Site Name NewPine Creek StatePark DevineCanyon Scenic Area Figure 4.10. Pre-release density/m2 of selected weevil release sites compared to 2013 plant survey.
49
100%
80%
60%
40%
20%
0%
-20% PercentChange -40%
-60%
-80%
-100%
OIT
Klatt
Tumalo
1815Rd
Mcallister
CellTower
Hwy140 W
GaslineRdII
Haglestein1
Skylinerpit 3
Harper Creek
ServiceCreek
DevineCanyon
GoldenPasture
Hwy140/coffee river
NewPine Creek StatePark DevineCanyon Scenic Area
ParadiseHills/Bald Mountain Release Site Name WillowCreek (beforeMiller's) Figure 4.11. Percent change from release density to 2013 density of Dalmatian toadflax. Mean reduction of Dalmatian toadflax density across all sites was 50%.
50
Dalmatian Toadflax Density Relative to Release Year
30
Individual Site Densities 25 Average Density
20
15
10 Density Density persquaremeter
5
0 0 1 2 3 4 5 6 7 8 9 10 11 12 Years Since Release
Figure 4.12. Change in density/m2 of Dalmatian toadflax relative to release year zero.
Mean density/m2 of Dalmatian toadflax at time of release (n = 20) was 9.45 ± 1.34, while sites monitored any time after release (n = 20) had an average of 5.5 ± 1.1.
Density of Dalmatian toadflax at sites prior to release of M. janthinus were found to be significantly lower than those after (Figure 4.12) (one tailed t-test = 1.68, p-value =
0.005). This indicates that biological control with M. janthinus is associated with reduction of the density of Dalmatian toadflax when compared to density at time of release. There was no significant difference found between individual years relative to release (Kruskal Wallis, p-value = 0.11).
51
Population Trends at Dufur Observation Site
The Dufur release site (Wasco County), was monitored over a 13 year period for abundance of M. janthinus, average density/m2 of Dalmatian toadflax, and percentage of toadflax infested by the weevil. Evidence of a predator-prey, density dependent host prey relationship was observed (Figure 4.13.a). After an initial failed establishment of M. janthinus in 2001, the 2003 release saw a successful establishment with a subsequent boom in population of M. janthinus in 2004. Prior to successful establishment of M. janthinus, the site had an estimated average
Dalmatian toadflax density of 8 plants/m2 (Figure 4.14.a). This rapid expansion of M. janthinus correlated to a precipitous decline in Dalmatian toadflax densities at the site, from 8 plants per square meter in 2004 to 0.01 in 2006 where M. janthinus had established on the toadflax (Figure 4.14.a). There was also a relationship present between the percent of plants infested by the weevil, and corresponding estimated average density of Dalmatian toadflax on the site (Figure 4.14.b). At low densities, high numbers of weevils per plant can occur from large numbers produced in prior years.
It was observed before M. janthinus had established across the entirety of the site in
2004 that the weevil was moving as a “biological wildifre” across the Dalmatian toadflax infestation, with a clear line of delineation where the weevil populations had generated heavy damage on their host (Figure 4.15). Within the bounds of control,
52 populations were at the aforementioned 0.01/m2, while outside the bounds of control populations remained between 8 - 15 plants/m2. By 2006, M. janthinus had spread throughout the site and reduced plant density to 0.01/m2 down from an estimated 8/m2 in 2004, representing a three order of magnitude reduction of toadflax density (Figure 4.14.a.). This pattern appeared to be a diffusion of the weevil through the toadflax infestation and may reflect spread rate of M. janthinus within a given host population.
At the Dufur site, establishment of a weevil population that reached high enough levels to reduce its host’s density levels was between 3 and 4 years, with the crash of toadflax populations occuring after levels of the weevil abundance reached a count of 100 per minute. After establishment of M. janthinus at Dufur site, the populations of both species begin a classic predator-prey dynamic population state (Abrams
2000). Although not measured in this study, it was observed that the resurgence of toadflax densities did not return to the original biomass and correlarily propagule production that was observed prior to release of M. janthinus. The spike in estimated average density that occurred from 2009 to 2011 is postulated to have significantly less ecological impacts due to the overall reduction in biomass and seed production of the toadflax (Appendix 4). These observations were made at numerous release sites where overall site biomass was significantly reduced.
53
9 120
8 100 7
2 6 80 /m 5 60 4 Density of Toadflax
Plant Plant Density 3 Weevil 40 Abundance 2 20 Weevil abundanceperminute 1
0 0 2001 2002 2003 2004 2006 2007 2008 2009 2010 2011 2013 Year of Observation Figure 4.13.a. Change in density/m2 of Dalmatian toadflax relative to abundance of M. janthinus.
54
9 120
8 100 7
2 6 80
5 60 4 Density of Toadflax
Plant Plant Density/m 3 40 % Plants Infested 2 20
1 Percentofplants infestedby M. janthinus
0 0 2001 2002 2003 2004 2006 2007 2008 2009 2010 2011 2013 Year of Observation
Figure 4.13.b. Change in density/m2 of Dalmatian toadflax relative to percent infestation of plants or damage by M. janthinus.
55
Figure 4.14.a. Dalmatian toadflax densities averaged 8/m2 at the Dufur site in 2003 prior to successful establishment of M. janthinus. Small tree used as reference point. Photo Credit: Eric Coombs, ODA.
56
Figure 4.14.b. Dalmatian toadflax densities crashed precipitously by 2006, three years after establishment by M. janthinus on the site. This picture from 2006 shows the negative relationship between M. janthinus and Dalmatian toadflax densities. Small tree in center used as reference point. Photo Credit: Eric Coombs, ODA.
57
Figure 4.15. “Biological wildfire” of M. janthinus moving its way through the toadflax infestation at the Dufur Site in 2004. Heavily damaged toadflax stalks from previous years delineate the area of control, while the standing toadflax shows area where M. janthinus is at low density or has yet to establish. Line used as demarcation of weevil impact. Photo Credit: Eric Coombs, ODA
58
Ecological Amplitude of Dalmatian Toadflax
Figure 4.16. The mean distribution of a habitat suitability model for Dalmatian toadflax and overlaid known weed locations.
The habitat suitability model generated in KRESS approximated the potential mean ecological amplitude of Dalmatian toadflax at 12,338,339 hectares (30,488,700 acres), or approximately 48% of the state land area (Figure 4.16). Current acres impacted by Dalmatian toadflax represent 1% of this potential mean biological potential. Utilizing ODA approximated current acreages of lands in Oregon impacted by Dalmatian toadflax at (ODA 2013) 142,835 hectares (325,955 acres), and the first
59 known infestation occurring in 1908 (Rice 2013), an annual expansion of 3% at 1,360 hectares (3,361 acres) has occurred since the introduction of the weed, assuming linear expansion. This is lower than the annual expansions rates of 14% and 6% of
BLM and National Forest lands in the Intermountain West (Dewey et al. 1995), respectively, and may be attributed to Dalmatian toadflax’s limited seed spread capacity as compared to other weedy species.
Using the mean of the habitat suitability model, intersections between resources impacted by Dalmatian toadflax and its potential range can be calculated. rangelands, right-of-ways, and riparian zones were deemed to the most threatened by Dalmatian toadflax infestations (ODA 2013, Figure 4.4.a). Rangeland resources were the most heavily impacted with 61%, or 4,859,474 hectares (12,008,023 acres) of range in Oregon impacted. Right-of-way had 13% of its overall structure impacted at 73,738 hectares (182,211 acres). Riparian zones had 42% of their statewide distribution impacted by toadflax at 362,520 hectares (895,808 acres).
Data pertaining to economic costs of Dalmatian toadflax infestations is limited
(Lajeunesse 1999), but the main impacts appear to be affiliated with direct management, reduced cattle-carrying capacity, and appraised value of infested land
(Lacey and Olsen 1991). Direct management costs averaged $99 per hectare ($40 per acre) on a Montana ranch of which Dalmatian toadflax exhibited 25-100% vegetative
60 cover (Lajeunesse 1999). A USDA Land Values 2012 Summary found that the 2012 statewide average annual value per acre for pasture land, which includes any land that is grazed by livestock, was $620 (USDA 2012). Management costs of Dalmatian toadflax could reduce this annual value per acre 6.5%. Because the USDA figure of
$620 average annual value per acre includes more productive Willamette Valley pasture within its calculation, more marginal grazing land east of the Cascade
Mountains with lower vegetative production could see larger impacts on a percentage basis to profitability per acre. The habitat suitability model predicts that
Dalmatian toadflax will be more prevalent east of the Cascade Mountains, so higher impacts to less productive land is likely.
A Receiving Operator Characteristic (ROC) analysis was used to assess the validity of the predicted model by comparing the suitability map to known occurrences of
Dalmatian toadflax, where a ROC value of 0.5 indicates randomness amongst the correlation between the distribution of Dalmatian toadflax and the predicted model, while higher values indicate a higher correlation between known Dalmatian toadflax infestations and the model (Harris 2001). The overall model received a 0.81 value from the ROC analysis. Of 4,889 Dalmatian toadflax locations, 4,495 points fell within the mean predicted zone which represented a 91% capture rate.
61
Chapter 5 – Discussion
This retrospective study is an attempt to chronicle the emerging success of the M. janthinus classical biological control program that was implemented by ODA, APHIS and their partners. There is evidence from the release site comparisons and the long-term observation site, that the presence of M. janthinus on Dalmatian toadflax infestations negatively impacts plant density (Figure 4.11 and 4.13.a). There was strong statistical evidence to reject the null hypothesis that densities of Dalmatian toadflax prior to, and after release of M. janthinus were equivalent (Figure 4.12). The
50% observed average net reduction in toadflax densities at release sites over time indicates that a measure of control has been achieved by this classical biological control program. With individual sites seeing reductions of toadflax density up to
98%, the potential for more robust control is evident. This initial reduction in toadflax densities should be considered in the context of this study’s relatively early analysis of a typical biocontrol program which takes 20-30 years to determine whether the program was successful (Coombs et al. 1996).
Because of the observational nature of this study, a regimented and repeated data collection protocol was not available with the exception of the long-term Dufur site.
Because of this, only snapshots of the population dynamic between predator-prey relationship were observed (Figure 4.9, 4.10 and 4.11). Expansions and collapses of
62 both species populations are cyclical in time, and appear to generally inversely track each other (Figure 4.13.a). If M. janthinus were having limited or no impact on
Dalmatian toadflax densities, one would expect the release sites to have a more equal distribution of increasing and decreasing toadflax densities. Yet 95% of the sites were decreasing relative to their density upon release. This indicates an overall declining trend that may represent an asymptotic trend that is not apparent in the
Dufur long-term study site (Figure 4.13.a). Although not found to be statistically significant between years, the toadflax density relative to release year shows this potential asymptotic trend (trend line included) (Figure 4.12). The small p-value observed at year four (p = 0.07) relative to year zero should also be considered through a biological interpretation of significance, where non-statistical significance may still contain biological relevance (Martı´nez-Abraı´n 2008). To confirm whether toadflax densities follow an asymptotic trend following a weevil introduction, data collection across multiple sites with substantial replicates would need to be implemented.
There was also evidence that M. janthinus has begun to spread beyond the original release sites independently of human facilitation to infestations of toadflax up to 60 kilometers away (Figure 4.8). Considering the ability of M. janthinus to fly, it is to be expected that a mobile herbivore would move toward more ample food sources as carrying capacity of toadflax infestations is reached. The more extreme values of
63 distance from the nearest release site may have resulted from accidental human transport, a non-reported release or wind-borne migration.
There were five sites in the 2013 survey that were more than 30 kilometers from the nearest release, with the exception of one site, the other four were within 1.5 kilometers of another Dalmatian toadflax population (ODA 2011). Given that the median distance from release site was 1.5 kilometers in the 2013 survey data, it is probable that weevils have been traveling from one infestation to the next across
Oregon, flying relatively short distances to incrementally expand their range. It is also feasible that wind-borne migration has facilitated the movement of M. janthinus across the state, with research showing that many species of insects use adaptive wind-borne migration to spread to new habitat (Gatehouse 1997). There may have been a correlation early in the M. janthinus release program between the abundance of weevil and distance from release sites, but the distribution program and the natural weevil migration have obscured any correlation.
The management implications of the natural long-distance dispersal of an impactful biological control agent should be considered in future release programs of different agents that have similar migration potential. The 219 known releases may represent an over-application of weevils, and a use of resources that could have been utilized for other biological control programs or weed control projects. With tightening
64 government budgets, establishing regionally successful “sources” by which the agent naturally migrates from to other regional infestations may represent a more cost- effective way of achieving successful biological control programs with reduced costs.
The control and spread capabilities of M. janthinus could support the grazing economy in Eastern Oregon which may be impacted if Dalmatian toadflax were left unchecked. The release form data shows that Dalmatian toadflax is primarily a rangeland weed in shrubland and grassland communities, where cattle production is often concentrated. At low densities, livestock can avoid the plant, and consequently its impacts are low. Conversely, at high densities cattle begin to be excluded from land once capable of being grazed. In addition to the potential 6.5% reduction in net value per acre, displacement of cattle via toadflax infestations, and thus reduced cattle productivity could result in areas impacted by Dalmatian toadflax.
Determining the economic impacts of Dalmatian toadflax to grazing production due to displaced cattle and reduced land values would assist in further rationalizing the long-term economic benefits of biological control.
Grazing appeared to be the economic activity that was most affiliated with infestations, whether or not it facilitates infestations would require further study.
Dalmatian toadflax only requires small disturbances to allow for establishment
(Lajeunesse 1999), and grazing may provide those open niches. To this end, the
65 introduction of M. janthinus has the potential to improve rangelands degraded by
Dalmatian toadflax over time.
Impacts of toadflax infestations to biodiversity of infested areas have not been thoroughly studied, but there is ample evidence that ecosystems can have reduced plant community diversity when invasive species are present (Schooler et al. 2006,
Groves and Willis 1999). Because of the wide range of habitats that Dalmatian toadflax can occupy in natural areas that are often difficult to access for conventional management, the reduction of toadflax densities via biocontrol provides a landscape scale management approach that results in less competition of toadflax on desirable plant communities. Further study in determining whether there is a net benefit to desirable plant communities by measuring replacement vegetation at release sites would help quantify the net impacts to biodiversity due to biocontrol of Dalmatian toadflax.
The average densities across different ecological, geophysical and land use types indicates that Dalmatian toadflax can establish and flourish across a wide ecological threshold, and shows the resiliency of the weed to different environments. Utilizing
M. janthinus as a control agent is the best response to Dalmatian toadflax’s wide ecological amplitude; the weevil has the ability to control the weed on a scale that is not economically viable through traditional control. ODA approximated that the
66 average cost per release of M. janthinus during the course of the program was $350 including labor, transport and monitoring (T. Butler, personal communication,
November 3, 2013). With a total of 219 reported releases, the total cost to the taxpayer equated to $76,650 over the course of the biocontrol program. If ODA were to have used conventional control with herbicides instead using that same amount of money, 1,916 acres would have been treated at $40 per acre out of an estimated
352,955 acres impacted by Dalmatian toadflax in Oregon. This equates to 0.05% of the current infestation, and only one year of treatment. When framed relative to the costs, spatial and temporal constraints of conventional control, the M. janthinus release program appears to be cost effective and represents a spatially holistic approach to toadflax management.
To better understand the natural migration patterns of M. janthinus, northern
California may prove to be an excellent testing ground for determining natural migration patterns of the weevil. It has varying levels of Dalmatian toadflax infestations, but has yet to approve M. janthinus for release. Because of its close in proximity to the weevil rich Dalmatian toadflax infestations of Klamath Falls, Oregon, the progression of the weevils over time from their Oregon sources could be monitored and population dynamics tracked. Any future study on the effect of a biological control agent on its host’s overall influence on a given ecosystem should include a metric more ecologically relevant than just density. Utilizing a metric such
67 as the importance value which measures the relative frequency, relative density and relative dominance of a species into a singular value better evaluates the impact of given species on ecosystem more accurately than density alone (Burns and Honkala
1965).
There are additional studies that have also found that the weevil is an effective biological control against Dalmatian toadflax (Nowierski 2004, De Clerck-Floate and
Miller 2002, Schat et al. 2011), in addition to it establishing regionally independent of human distribution (Van Hezewijk 2010). From the evidence presented, there appears to be no ecological or financial incentive for ODA, APHIS and their partners to continue M. janthinus releases. Funding for further monitoring of ODA release sites would provide valuable information about population dynamics, and give further validation of program success. From all the evidence provided, it appears that the weevil will naturally continue to migrate to new infestations of Dalmatian toadflax, reach sufficient carrying capacity population levels for control, and migrate again to new infestations thus continuing its trend of expansion and inhibition of its target species: Dalmatian toadflax.
68
Bibliography
Abrams, P. A. (2000). The evolution of predator-prey interactions: Theory and evidence. Annual Review of Ecology and Systematics, 31, 79-105.
Alex, J. F. (1962). The taxonomy, history, and distribution of Linaria dalmatic. Canadian Journal of Botany, 40, 295-307.
Arnold, W.R., & Santelmann, P.W. (1966). The Response of Native Grasses and Forbs to Picloram. Weed Science Society of America. 14(1), 74-76.
Breiter, N. C., & Seastedt, T. R. (2007). Postrelease evaluation of Mecinus janthinus host specificity, a biological control agent for invasive toadflax (Linaria spp.). Weed Science, 55(2), 164-168.
Burns, R. M., & Honkala, B. H. United States Department of Agriculture, United States Forest Service (1965). Silvics of North America (86-60058). Retrieved from Oxford Press website: http://www.na.fs.fed.us/spfo/pubs/silvics_manual/table_of_contents.htm
Butler, T. (2013, November 3). Personal Communication.
Coombs, E.M., (2013, December 1). Personal Communication.
Coombs, E.M. (2013). Biological control of weeds in Oregon. Annual Report. Oregon Department of Agriculture. Salem, OR. 41 pp.
Coombs, E.M., Radtke, H., Isaacson, D.L., Snyder, S.P. (1996). Economic and regional benefits from the biological control of tansy ragwort, Senecio jacobaea, in Oregon. Proceedings of the IX International Symposium on Biological Control, 489- 494.
Daly, C. & Taylor, G. (2001). United States Average Monthly and Annual Climate Data, 1971-1990. The Climate Source, Inc.
De Clerck-Floate, R., & Miller, V. (2002). Overwintering mortality of and host attack by the stem-boring weevil, Mecinus janthinus germar, on Dalmatian toadflax (Linaria dalmatic (l.) mill.) in western Canada. Biological Control, 24(1), 65-74.
Dewey, S.A., Jenkins, M.J., and Tonioli, R.C. (1995). Wildfire suppression - a paradigm for noxious weed management. Weed Technology, 9(3), 621-627.
Gatehouse, A. G. (1997). Behavior and ecological genetics of wind-borne migration by insects. Annual Review of Entomology, 42, 475-502.
69
Goulet, E. J., Thaler, J., Ditommaso, A., Schwarzländer, M., & Shields, E. J. (2013). Impact of Mecinus janthinus (coleoptera: Curculionidae) on the growth and reproduction of Linaria dalmatic (Scrophulariaceae). Great Lakes Entomologist, 46(1- 2), 90-98.
Groves, R.H., & Willis, A.J. (1999). Environmental weeds and loss of native plant biodiversity: Some australian examples.Australasian Journal of Environmental Management, 6(3), 164-171.
Harris, N. (2001). Cattle behavior and distribution on the san joaquin experimental range in the foothills of central California. (Doctoral dissertation).
Hijmans, R.J., Cameron, S.E., Parra, J.L., Jones, P.G., & Jarvis, A. (2005). Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology, 25, 1965-1978.
Jacobs, J. S., & Sheley, R. L. (2003). Prescribed fire effects on dalmation toadflax. Journal of Range Management, 56(2), 193-197.
Jeanneret, P., & Schroeder, D. (1992). Biology and host specificity of Mecinus janthinus germar (Coleptera: Curculionidae), a candidate for the biological control of yellow and dalmatian toadflax, Linaria vulgaris (L.) Mill. and Linaria dalmatica (L.) Mill. (Scrophulariaceae) in North America. Biocontrol Science & Technology, 2(1), 25- 34.
Johnson, D.E. (2013, May 10). In-person interview.
Johnson, D.E., Louhaichi, M., Harris, N.R., Clark, P.E., Ganskopp, D., Johnson, M.D., Johnson, K.E., George, M.R. (2005) a. Kress modeler version 4.0.0 user’s manual kress project. Corvallis, OR: Oregon State University.
Johnson, D.E., Louhaichi, M., Harris, N.R., Clark, P.E., Ganskopp, D., Johnson, M.D., Johnson, K.E., George, M.R. (2005) b. KRESS Modeler Tutorial Exercises. Corvallis, OR: Oregon State University.
Lacey, J., & Olsen, B. (1991). Environmental and Economic Impacts of Noxious Range Weeds. James, L.F. et al. (Eds.) In Noxious Range Weeds (5-16). Boulder, CO: Westview Press.
Lajeunesse, S. (1999). Dalmatian and Yellow Toadflax. In: Sheley, R.L., Petroff J.K., (Eds.), Biology and Management of Noxious Rangeland Weeds (202-216). Corvallis, OR: Oregon State University Press.
70
Lajeunesse, S. E., Fay, P. K., Cooksey, D., Lacey, J. R., Nowierski, R. M. and Zamora, D. Montana State University Extension Service, (1993).Dalmatian and yellow toadflax: weeds of pasture and rangeland. Bozeman, MT.
Mack, R. N. (2003). Plant naturalizations and invasions in the Eastern United States: 1634–1860. Ann. Mo. Bot. Gard. doi:10.2307/3298528
Martı´nez-Abraı´n, A. (2008). Statistical significance and biological relevance: A call for a more cautious interpretation of results in ecology. Acta Oecologica, 34, 9-11.
McClay, A. S., & Hughes, R. B. (2007). Temperature and host-plant effects on development and population growth of Mecinus janthinus (Coleoptera: Curculionidae), a biological control agent for invasive Linaria spp. Biological Control, 40(3), 405-410.
Nowierski, R.M. (2004). Toadflaxes. In: Coombs, E.M., Clark, J.K., Piper, G.L., and Cofrancesco A.F. (Eds.), Biological Control of invasive plants in the United States (392- 294). Corvallis, OR: Oregon State University Press.
Oregon Department of Agriculture. (1983). Unpublished Weed List. Oregon Department of Agriculture. Salem, OR.
Oregon Department of Agriculture. (2011). Weedmapper. Retrieved from www.weedmapper.Oregon.gov/
Oregon Natural Heritage Information Center. (2010). Ecological System Life Form Data. Portland, OR.
Rasmussen, P. (2013, November 3). Email.
Rice, P. (2013, January 7). Invaders database system. Retrieved from http://invader.dbs.umt.edu/
Rinella, M.J. Maxwell, B.D., Fay, P.K., Weaver, T., and Sheley, R.L. (2009). Control effort exacerbates invasive-species problem. Ecological Applications, (19), 155-162.
Robocker, W. C. (1970). Seed characteristics and seedling emergence of dalmatian toadflax. Pullman, WA: Washington Agricultural Experiment Station.
Robocker, W. C. (1974). Life history, ecology, and control of dalmatian toadflax. Pullman, WA: Washington Agricultural Experiment Station.
Seaber, P.R., Kapinos, F.P., and Knapp, G.L., (1987). Hydrologic unit maps: U.S. geological survey water-supply. United States Geological Survey.
71
Schat, M., Sing, S. E., Peterson, R. K. D., Menalled, F. D., & Weaver, D. K. (2011). Growth inhibition of Dalmatian toadflax, Linaria dalmatic (l.) miller, in response to herbivory by the biological control agent Mecinus janthinus Germar. Journal of Entomological Science, 46(3), 232-246.
Schooler, S.S., McEvoy, P.B., and Coombs, E.M. (2006). Negative per capita effects of purple loosestrife and reed canary grass on plant diversity of wetland communities. Diversity and Distributions, 12, 351-363.
Sing, S. E., & Peterson, R. K. D. (2011). Assessing environmental risks for established invasive weeds: Dalmatian (Linaria dalmatic) and yellow (L. vulgaris) toadflax in north america. International Journal of Environmental Research and Public Health, 8(7), 2828-2853.
Thomas, M.B., & Adele R.M. (2007). Are exotic natural enemies an effective way of controlling invasive plants? Trends in Ecology and Evolution, 22(9), 447-453.
Toševski, I., Caldara, R., Jović, J., Hernández-Vera, G., Baviera, C., Gassmann, A., & Emerson, B. C. (2011). Morphological, molecular and biological evidence reveal two cryptic species in Mecinus janthinus germar (coleoptera, curculionidae), a successful biological control agent of dalmatian toadflax, Linaria dalmatic (lamiales, plantaginaceae). Systematic Entomology, 36(4), 741-753.
Toševski, I., Jović, J., Krstić, O., & Gassmann, A. (2013). PCR-RFLP-based method for reliable discrimination of cryptic species within Mecinus janthinus species complex (mecinini, curculionidae) introduced in north america for biological control of invasive toadflaxes. Biocontrol, 58(4), 563-573.
United States Department of Agriculture, National Agricultural Statistics Service (2012). Land values 2012 summary (ISSN: 1949-1867). Retrieved from website: http://www.nass.usda.gov/Publications/Todays_Reports/reports/land0812.pdf
United States Department of the Interior, United States Geological Survey. (2012). Global 30 arc-second elevation (gtopo30). Retrieved from website: https://lta.cr.usgs.gov/GTOPO30
Van Hezewijk, B. H., Bourchier, R. S., & De Clerck-Floate, R. A. (2010). Regional-scale impact of the weed biocontrol agent Mecinus janthinus on Dalmatian toadflax (Linaria dalmatica). Biological Control, 55(3), 197-202.
Vujnovic, K., & Wein, R. W. (1997). The biology of canadian weeds. 106. Linaria dalmatica (L.) mill. Canadian Journal of Plant Science, 77(3), 483-491.
72
Wilson, L.M., Sing, G.L., Piper, R.W., Hansen, R. De Clerck-Floate, D.K. MacKinnon, and Randall, C. (2005). Biology and biological control of Dalmatian and yellow toadflax. USDA Forest Service, FHTET-05-13.
73
Appendices
Appendix 1 – Biological Control Agent Release Form
74
75
Appendix 2 – Biological Control Monitoring Form
76
Appendix 3 – Dalmatian Toadflax Population Trends at Release Sites
Release Site Name Release 2013 Percent Years Since Density Density Density Change Release Trend (Density (Density per square per meter) square meter) Golden Pasture 8.5 1 -0.12 11 Declining
Devine Canyon 8.5 1 -0.12 11 Declining
OIT 18 10 -0.56 11 Declining
Devine Canyon Scenic Area 8.5 2.5 -0.29 10 Declining
Harper Creek 5 2 -0.40 8 Declining
Hwy 140/coffee river 18 15 -0.83 7 Declining
Paradise Hills/Bald Mountain 18 5 -0.28 7 Declining
Tumalo 6 5 -0.83 6 Declining
Skyliner pit 3 4.5 0.1 -0.02 6 Declining
Gasline Rd II 8 15 +0.88 6 Increasing
Hwy 140 W 3 2 -0.67 6 Declining
Service Creek 10 6 -0.60 6 Declining
1815 Rd 26 15 -0.58 5 Declining
Willow Creek (before Miller's) 10 5 -0.50 5 Declining
Cell Tower 10 1 -0.10 5 Declining
Mcallister 6 2.5 -0.42 4 Declining
Klatt 7 2.5 -0.36 4 Declining
Haglestein 1 6 2.5 -0.42 4 Declining
New Pine Creek State Park 5 3 -0.60 4 Declining
77
Appendix 4 – Impacts of Non-biocontrol Management on Dalmatian Toadflax Populations Infested with Mecinus janthinus
A Burns, Oregon infestation on a vacant lot that has been repeatedly mowed, reducing M. janthinus population densities to a level where sufficient control of the toadflax has been negated. Biomass and density are still high, and represents juxtaposition to the trend across sites where the weevil is unhindered in reproducing to a population density levels capable of controlling the weed.
78
Tumalo release site where estimated average plant density levels had resurged from previous years of control. Note that although density may still be relatively high, biomass has been significantly reduced.