Transcript Document

Interplay between contamination and the risk of invasive species or genetically modified organisms at population and landscape scales.
Wayne G. Landis, Audrey M. Colnar, Ananda Seebach and Laura J. Sellens, Institute of Environmental Toxicology, Huxley College of the Environment, Western Washington University, Bellingham WA, USA
Abstract
2
In many ways the factors governing the establishment of invasive species and genetically
modified organisms (GMOs) are closely related topics. The introduction and establishment of
invasive species or genetically modified organisms appears to be dependent upon the
characteristics of the individuals, the life history strategies of the invasive, the pattern of
habitats in the landscape and other stressors in the environment. Modeling by Deines, Chen
and Landis (2005) has demonstrated that contaminants can influence the likelihood of
introduction by affecting competitors, changing rates of migration to other habitats or altering
the fitness of the invasive. Changes in the arrangement and size of habitat patches via
indirect effects can have as large an effect as any direct effect when the number of and
location of beachheads are changed. In the case of GMO bentgrass for herbicide resistance,
the introduction of the specific herbicide can alter the selection landscape, allowing hybrids a
selective advantage and increasing the probability that herbicide resistant strains of
hybridizing species could result. In summary, contaminants may discourage or encourage the
establishment of invasives or GMOs by altering the selective landscape at multiple biological
scales.
Figure 3. Basic framework for risk assessment incorporating
invasive species
Stages of
General Model
Invasion
Integrated Hierarchical Conceptual Models
Introduction at the
Local Scale
Local Scale
We have incorporated the HPDP into the RRM framework
so that all hierarchies of scale are integrated with the risk
assessment process (Figure 3). Audrey Colnar has
applied the RRM-HPDP process to the risk assessment
for the European Green Crab (Colnar and Landis 2007).
The hierarchical approach allowed organization of the
conceptual model by both scale and risk assessment
process so that risk could be calculated. The effect of
contaminants upon the invasives can be incorporated at
each of these scales. Uncertainty is estimated using
Monte Carlo analysis. This same approach has been
used to estimate risks due to the Nun Moth, the Asian
Oyster and for genetically modified organisms.
Habitat patches within
Cherry Point
Focal Scale
Patchy populations
within the habitat
patches within Cherry
Point
Colonization at the
Local Scale
Establishment at the
Local Scale
Dispersal (via
reproduction) and
Migration
Regional Scale
Large-scale patches
beyond Cherry Point
Invasive Spread at the
Focal Scale
4
Application-Genetically Modified Organisms-Sellens and Landis
Many aspects of the spread of genetically modified organisms are similar to invasive species. In the case of
GMOs the gene can be the marker that is being considered. In this instance we modeled a risk scenario based
upon a herbicide resistant strain of bentgrass near Madras, Oregon. Creeping bentgrass (Agrostis stolonifera L.)
is a high value golf course turf grass. It differs from previously deregulated species in having a stoloniferous
growth habit, and in having at least 13 native and non-native species with which it can freely hybridize. In 2003,
eight fields totaling approximately 162 ha of RoundUp Ready® creeping bentgrass were grown within a 4,453 ha
control district outside of Madras, Oregon. Watrud et al. (2004) found pollen-mediated gene flow up to 21 km
from this site, about 10 times the expected distance. Reichman et al. (2006) has demonstrated that gene flow
has occurred to surrounding populations.
Figure 8 presents the conceptual model at the -1 scale for the bentgrass that incorporates glyphosate
application. This stressor is one of three that form a selective pressure on the host.
Habitat Suitability
Proximity of
Compatible
Plants
Invasive Spread at the
Regional Scale
1
Size of
Source
Patch
Background and Theory
Water
Flow in
Canals
3
Application-Invasives
S
T
A
R
T
Three risk assessments have been performed for a
number of endpoints at Cherry Point, Washington (Figure
4), one dealing with contaminants and physical stressors
and two with invasive species. The methods used were
comparable, including the Relative Risk Method and
similar processes of ranking. The results are summarized
and compared below.
Human Intentional Introductions
Ornamentals
Crops
GEOs
Biocontrol organisms
Recreational introductions
Habitat types
Marine near shore
Estuaries
Rivers and streams
Agricultural areas
Urban Parks
Commercial areas
Industrial areas
Forests
Natural Introductions
Species shifts and expansions due to
climate
change
Colonization due to natural events
(storms,
currents, winds)
Pathogens (chestnut blight, West Nile
virus, HIV
Wildlife species (pheasant, house
sparrow, starlings, green crab,
mongoose)
Plant species (kudzu, Spartina)
Insect pests
Animal
Transport
Pollen
Viability
1400
Impact types
Replacement of economically,
culturally or socially important
species
Decrease in available habitat
for significant species
Decrease in biodiversity
Transformation of habitat type
Disease (human. plant and
wildlife)
Change in population
dynamics
Alterations of landscape
structure
Cherry Point
T otal Ris k S c ore 2206
P oint Roberts
T otal Ris k S c ore 1449
1200
Control
District
Pollen
Flow
Seed
Stolon
Plug
Mowing
Human
Transport
Backcrossing
Duration
of Seed
Selective
Pressure
Nonexpression of
gene
Biological
Parameters
Competitive
Interactions
Machinery
Transport
Anthropogenic
Parameters
Controlling
Factors
HABITATS
EFFECTS
Grassland
Pollinization/
Hybridization
Canals
Seedling
Growth
Agricultural/
Range
Vegetative
Translocation
IMPACTS
Extinction of
Plant
Establishment
of Plant
SECONDARY
SOURCE OF
GENE
Local scale (-1) of the hierarchical conceptual model. This scale shows the pathways by
which the gene could travel at the individual level. The solid lines show movement via
pollen, the dashed line shows movement through seeds, and the dash-dot line shows
movement through vegetative escape. The factors influencing movement for each of these
pathways are shown. The point influenced by the herbicide application is in red.
1600
Effect
PATHWAYS OF
STRESSOR
Pollen Size
Synchrony of
Compatible
Plants
Glyphosate
Application
Grazing
Urban
Relativ e risk to biological assessment endpoints in sub-regions
Impact
Stressor
Wind
Speed
Seed Size
Sagebrush
Figure 5. Chemical and Physical Stressors (Hart Hayes and Landis 2004)
Human Accidental Introductions
Tankers (Ballast water and hulls)
Trucking (Cargo, plants and animals)
Aircraft ( On clothing, foodstuffs)
Wind
Direction
Physical
Parameters
SOURCE
Figure 1. Invasive species and Regional Risk Assessment
Habitat
Location of
Compatible
Plants
Figure 4. Cherry Point Washington study area
In order to estimate risk we base our approach upon the regional model developed by Landis
and Wiegers (1997) that incorporates sources, stressors, habitats, effects and impacts (Figure
1). The relative risk model (RRM) has been adapted to the estimate of risk for invasive
species (Landis 2004). The sources correspond to the areas or transportation activities from
which invasives are derived. The stressors are the invasives. Habitats correspond to the
specific locations and areas that are colonized by the introduced species. Interactions with
other species and the physical attributes of the ecological system generate a series of effects
that when they negatively affect the valued components result in impacts.
Source
Toxicant Interaction
GMO Conceptual Model
Figure 6. European Green Crab (Colnar and Landis 2007)
Figure 9. Local level diagram of gene flow and establishment
of resistant plants.
1000
800
600
Figure 10. Total risk estimate for
each assessment endpoint.
400
200
0
CS
DC
ES
GBH
LC
SSE
CS
DC
ES
GBH
LC
SSE
LC
SSE
1600
Lummi
T otal Ris k S c ore 3564
Drayton Harbor
T otal Ris k S c ore 3072
1400
1200
1000
Mortality due to disease or parasitism
Predation
Interspecific competition
Intraspecific competition
Alteration of the physical habitat
Alteration of nutrient pathways
Change in migration matrix
800
Landis et al, 2004
Risk Analysis
600
400
200
0
CS
DC
ES
GBH
LC
SSE
CS
DC
ES
GBH
1600
B irc h Bay
T otal Ris k S c ore 2372
1400
A lden B ank
T otal Ris k S c ore 228
1200
One of the biggest difficulties in estimating risk in ecological systems is the problem of
dealing with multiple scales. We have adopted the Hierarchical Patch Dynamics
paradigm (HPDP) as developed by Wu and colleagues (Wu and Loucks 1995, Wu and
David 2002). This approach incorporates a variety of scales into an organizational
construct that allows the description of the invasive-risk process at scales from the
individual to the landscape. This is described in the next part of the presentation.
Figure 2. Hierarchical Patch Dynamics paradigm (HPDP)
The HPDP allows the incorporation of different layers of scale with a framework for the
interactions at each level (Wu and David 2002)
1000
Risk to the endpoint species can be
calculated for the region under study.
800
600
400
200
0
CS
DC
ES
GBH
LC
SSE
Relative risk to assessment endpoints in sub-regions. Y-axis is the
relative risk score. X-axis from left to right: CS = Coho salmon, DC =
juvenile Dungeness crab, ES = juvenile English sole, GBH = great
blue heron, LC = native littleneck clam and SSE = surf smelt
embryos. The contaminants or physical stressors did not enhance
any of the chosen endpoints.
Unlike contaminants, the European Green Crab had risks to
some endpoints, but would enhance others. The
enhancement is due to the crab acting as a food source.
These risks are for an El Nino year when the currents would
be more likely to spread the European Green Crab from
coastal Washington.
The top portion shows the classic stressor-receptor-impact flow
chart for environmental risk assessment and the bottom portion
shows how the RRM can be applied directly to GM bentgrass.
5
Level +1
Higher Level
Context,constraints,
control, containment,
boundary conditions
Figure 7. Sargassum muticum (Seebach et al.)
Relative risk from S. muticum to each assessment
endpoint. Positive values indicate risk while negative
values indicate potentially beneficial effects on the
endpoints. In the case of Sargassum the habitat it
provides results in negative scores for many of the
endpoints.
Coho Salmon
Eelgrass
Level -1
Focal Level
Lower Level
Dynamic
Interactions
Components,
mechanisms, initial
conditions
Vertical Structure-Asymmetric relationships, Loose vertical coupling, a variety of ordering
principals.
Horizontal Structure-Symmetric relationships, Loose horizontal coupling, a variety of
strengths of interactions between components.
Assessment Endpoint
Great blue heron
Level 0
Conclusions
Juvenile Dungeness crab
Juvenile English sole
Littleneck clam
Macroalgae
Pacific herring
Surf smelt
-300
-200
-100
0
100
200
300
Risk Score
The invasive species can provide habitat or a food resource for a number of endpoints. Subsequent assessment with Asian Oyster
provided a similar pattern of both risk and enhancement. Nun moth has been the exception. The alteration of forests affected a
number of endpoints at a large scale.
1. It is possible to incorporate contaminant stressors into a risk assessment evaluation with
invasive species and genetically modified organisms.
2. The patterns of risk expressed by invasives tend to be different than from contaminants
towards typical endpoints.
3. In the evaluation of GMOs contaminants can be incorporated as a selective pressure
towards both the GMO and the organisms receiving the genetic construct.