Comparison of Two Milkweed (Asclepias) Sampling Techniques on Eastern Nebraska Grasslands
Mercy L. Manzanares1, Melissa J. Panella2*, Carissa L. Wonkka3,
Gerald A. Steinauer2, and Kristal J. Stoner4
1HDR, Inc., 1917 S 67th St., Omaha, NE 681062 USA. 2Nebraska Game and Parks Commission, 2200 N. 33rd St., Lincoln, NE 68503 USA. 3Agricultural Research Service, US Department of Agriculture, 1500 N. Central Ave., Sidney, MT 59270 USA. 4Audubon Nebraska, P.O. Box 22521, Lincoln, NE 68542 USA. *Corresponding author.
Praire Naturalist, Special Issue 1 (2022):54–64
Abstract
In recent years, interest in estimating the number of Asclepias Linnaeus (Milkweeds) on the landscape has grown, because the plants serve as hosts for the larvae of Danaus plexippus Linnaeus (Monarch Butterflies). Monarchs are a species of high conservation concern, whose status for federal listing under the Endangered Species Act is currently warranted but precluded by work on higherpriority listing actions. The widespread loss of milkweed throughout the Monarch’s range is one of the primary factors implicated in the butterfly’s decline. In 2017 and 2018, we led a study to compare estimates of milkweed abundance and richness from two sampling methods, belt-transects and plot sampling, in four land categories on 48 grasslands in eastern Nebraska. The four site types included: 1) high-diversity local ecotype plantings, 2) privately owned agricultural lands with grassland enhancement, 3) public wildlife management area grassland rehabilitations, and 4) unmanaged pastures (i.e., with no specific conservation management activities or plans). Overall, twice as many milkweed species and 31.80 times as many stems of milkweed were detected by the belt-transect method than by the plot method. More distinct differentiation was indicated by the transect method that was not captured by plot sampling with similar time and effort expended in the field. High-diversity planting sites and wildlife management areas had more milkweed species per hectare than unmanaged pastures and privately owned agricultural lands with grassland enhancement. Variance was much higher for counts obtained via plot sampling relative to counts from transect surveys. The findings from this study suggest that the belt-transect method is more effective than the plot method for predicting milkweed stems per hectare as well as milkweed species richness while expending similar effort in the field. The ability to choose the most useful methods for identifying and quantifying milkweed on a variety of land management types is paramount for implementing a conservation plan for the Monarch butterfly not only in Nebraska but throughout the midwestern United States.
Download Full-text pdf
54
1HDR, Inc., 1917 S 67th St., Omaha, NE 681062 USA. 2Nebraska Game and Parks Commission, 2200
N. 33rd St., Lincoln, NE 68503 USA. 3Agricultural Research Service, US Department of Agriculture,
1500 N. Central Ave., Sidney, MT 59270 USA. 4Audubon Nebraska, P.O. Box 22521, Lincoln, NE
68542 USA. *Corresponding author: melissa.panella@nebraska.gov
Associate Editor: Clint Otto, US Geological Survey, Northern Prairie Wildlife Research Center.
Comparison of Two Milkweed (Asclepias) Sampling
Techniques on Eastern Nebraska Grasslands
Mercy L. Manzanares1, Melissa J. Panella2*, Carissa L. Wonkka3,
Gerald A. Steinauer2, and Kristal J. Stoner4
Abstract - In recent years, interest in estimating the number of Asclepias Linnaeus (Milkweeds) on the
landscape has grown, because the plants serve as hosts for the larvae of Danaus plexippus Linnaeus
(Monarch Butterflies). Monarchs are a species of high conservation concern, whose status for federal
listing under the Endangered Species Act is currently warranted but precluded by work on higherpriority
listing actions. The widespread loss of milkweed throughout the Monarch’s range is one of
the primary factors implicated in the butterfly’s decline. In 2017 and 2018, we led a study to compare
estimates of milkweed abundance and richness from two sampling methods, belt-transects and plot
sampling, in four land categories on 48 grasslands in eastern Nebraska. The four site types included: 1)
high-diversity local ecotype plantings, 2) privately owned agricultural lands with grassland enhancement,
3) public wildlife management area grassland rehabilitations, and 4) unmanaged pastures (i.e.,
with no specific conservation management activities or plans). Overall, twice as many milkweed species
and 31.80 times as many stems of milkweed were detected by the belt-transect method than by the plot
method. More distinct differentiation was indicated by the transect method that was not captured by
plot sampling with similar time and effort expended in the field. High-diversity planting sites and wildlife
management areas had more milkweed species per hectare than unmanaged pastures and privately
owned agricultural lands with grassland enhancement. Variance was much higher for counts obtained
via plot sampling relative to counts from transect surveys. The findings from this study suggest that the
belt-transect method is more effective than the plot method for predicting milkweed stems per hectare as
well as milkweed species richness while expending similar effort in the field. The ability to choose the
most useful methods for identifying and quantifying milkweed on a variety of land management types
is paramount for implementing a conservation plan for the Monarch butterfly not only in Nebraska but
throughout the midwestern United States.
Introduction
Over the past two decades, it is estimated that the eastern population of Danaus plexippus
Linnaeus (Monarch Butterfly) has declined by as much as 84% (Brower et al. 2012,
Semmens et al. 2016, Vidal and Rendón-Salinas 2014). Because Monarch caterpillars forage
solely on Asclepias Linnaeus (milkweeds) in the Apocynaceae family, the abundance
of milkweeds in the Monarch’s upper midwestern breeding grounds is thought to be critical
for the species’ survival (Brower 1984, Pleasants and Oberhauser 2012). Several studies
have indicated a strong correlation between the amount of milkweed in breeding grounds
in the midwestern United States and overwintering Monarch populations in the central
Mexican Highlands (Brower et al. 2012, Flockhart et al. 2014, Pleasants and Oberhauser
2012, Stenoien et al. 2016). Reductions in the abundance of milkweed within the Monarch’s
breeding range have been attributed to land-use change (Pleasants 2017). Restoring eastern
Monarch butterfly populations will require establishing, rehabilitating, and managing
Pollinators of the Great Plains: Ecology, Distribution, and Conservation
2022 PRAIRIE NATURALIST Special Issue 1:54–64
Prairie Naturalist
M. L. Manzanares, M. J. Panella, C. L. Wonkka, G. A. Steinauer, K. J. Stoner
2022
55
Special Issue 1
habitat by using techniques that increase the abundance of milkweed across many landmanagement
types within grassland ecosystems.
In 2016, the Nebraska Monarch and Pollinator Initiative set a goal of establishing 125
million milkweed stems across the state. In order to reach this goal, it is necessary to understand
how different land-management types affect the establishment of milkweeds on a
variety of grassland landscapes. Asclepias is a genus of herbaceous perennial, dicotyledonous
plants native throughout a large portion of North America, excluding Alaska and parts
of the Pacific Northwest (Borders and Lee-Mäder 2014); at least 72 milkweed species, not
including subspecies, are native to the continental United States and Canada (Borders and
Lee- Mäder 2014). Some milkweed species occur as scattered stems, others form colonies
in dense stands, and some form clumps from a thick woody base. In the midwestern
US, milkweeds grow in a variety of habitats including prairies, woodlands, wetlands, road
rights-of-way, and croplands (Hartzler and Buhler 2000).
It is estimated that milkweed abundance declined 58% over the midwestern landscape
between 1999 and 2010 (Pleasants and Oberhauser 2012). This represents a loss of 861
million milkweed stems, with nearly two million additional milkweeds being lost annually
because of continued grassland conversion to other uses (Pleasants 2017). It has been
stated that to slow or halt the decline of Monarchs, significant habitat restoration must
occur (Nabhan et al. 2015). New populations of milkweed must be established and existing
populations enhanced within the grassland regions of the Monarch’s eastern breeding
grounds. The Mid-America Monarch Conservation Strategy (MAFWA 2018), for example,
seeks to add over one billion stems of milkweed in various habitats within the Monarch’s
midwestern breeding range. In the future, as conservationists, farmers, utility and roadside
managers, rural communities, and individuals restore Monarch habitat, there will be a need
to document changes in populations of milkweed to determine if conservation objectives are
met. Thus, it is necessary to evaluate different methodologies to assess milkweed abundance
and species richness.
The two objectives of this study were to compare two methods of estimating the abundance
and richness of milkweeds to determine their relative usefulness on sites in eastern
Nebraska, and to compare milkweed density across land categories. Understanding how
milkweed richness, abundance, and density change over time in response to habitat manipulation
is necessary for effective conservation decision-making.
Thogmartin et al. (2017) states the need for accurate estimates of milkweed based on
field studies across land cover types. The results from our study can better inform objectives
and actions in existing Monarch conservation plans in Nebraska (Panella 2017) and
the Midwest (e.g., Mid-America Monarch Conservation Strategy [MAFWA 2018]), as well
as guide updates to those plans and development of new conservation strategies.
Methods
The study area included the Monarch’s primary breeding range in Nebraska, which
corresponds to the tallgrass prairie ecoregion as described in the Nebraska Natural Legacy
Project (Schneider et al. 2011). Our field sampling occurred between 26 June to 6 October
2017, and from 29 May to 3 October 2018. We surveyed 48 grassland areas in four site types
by using two sampling methods, plots and belt-transects. The site types were distinguished
by the land-management practices that had been applied. These included high-diversity,
local-ecotype prairie plantings; wildlife management areas; Pheasants Forever restored agricultural
properties; and unmanaged pastures.
Prairie Naturalist
M. L. Manzanares, M. J. Panella, C. L. Wonkka, G. A. Steinauer, K. J. Stoner
2022
56
Special Issue 1
Site Types
Four different site types were surveyed during the summers (June to August) of the
study in a quasi-random order based on their geographic location and to accommodate for
efficient surveyor travel time among sites.
High-diversity, Local-Ecotype Plantings. Prairie Plains Resource Institute, Inc. facilitated
access to survey high-diversity sites that had been planted previously (2007–2014)
with native grasses and forbs, including Asclepias species. Seventeen properties were sampled
with an average site size of 11.50 ha. Prior to seeding, all of the properties had been
cropland. Most of these properties were on private land, and management differed depending
on the landowner. However, on most of the properties, land management involved a
combination of burning, grazing, or resting, and in some cases tree removal or spraying for
noxious weeds. Additionally, these sites were located in varying soils and across a moisture
gradient. All properties were planted with a high diversity of prairie plant species and a
milkweed mix, which included A. syriaca Linneaus, A. sullivantii Engelman ex A. Gray, A.
verticillata L., and A. speciosa Torrey, in addition to A. incarnata L. and A. arenaria Torr
depending on the site specifics. The seed mixes were local ecotypes and consisted of up to
100 different species of native grasses, sedges, legumes, composites, and other forbs.
Pheasants Forever Restored Agricultural Properties. Access to 13 privately owned
properties was facilitated by the Nebraska Pheasants Forever conservation program. These
properties were restored agricultural land that included corn and soybean field pivot corners
as well as more extensive tracts of land bordering agricultural production areas. These
sites had been planted between 2013 and 2014 with a seed mix that included two Asclepias
species, A. syriaca with either A. incarnata or A. arenaria. The average site size that we
sampled for these properties was 5.45 ha.
Unmanaged Pastures. Eleven unmanaged grassland pastures were included to provide
information regarding the “standard condition” of grazed, native pastures in Nebraska’s
tallgrass prairie ecoregion. These properties include land that may have supported crops
previously and currently have no specific conservation management activities or plans. Data
were collected from Common Land Units from the US Department of Agriculture’s Farm
Service Agency database (USDA FSA 2021). “Pasture” land type was randomly selected,
and then aerial imagery was used to select sites that were currently pasture. The average site
size of unmanaged pastures was 10.11 ha.
Wildlife Management Areas. Seven wildlife management areas were surveyed. These
state-owned properties are rehabilitated grasslands managed by the Nebraska Game and
Parks Commission. The properties selected were chosen because they had been managed
for wildlife habitat by using prescribed burning, noxious weed control, and grazing. The
average size of wildlife management areas was 59.6 ha.
Sampling Methods
Two sampling methods (i.e., belt transect and plot) were implemented on each site for
comparison of their effectiveness in detecting milkweeds. Both sampling methods occurred
in each field on the same day so as to avoid variations resulting from sampling date as new
stems continue to emerge and become more visible. Belt-transect sampling was conducted
first, followed by the plot method. Prior to field sampling, we used aerial photographs
and ArcMap 10.5 (Esri, Redlands, California, USA) to randomly place 10 or more evenly
spaced transects across a site; transects were aligned in a north-to-south or east-to-west
orientation and at least 10 m (32.8 ft) apart. The length of each transect varied from 100
m (0.06 mi) to a maximum of 2,000 m (1.2 mi) because of the irregular shapes of some
Prairie Naturalist
M. L. Manzanares, M. J. Panella, C. L. Wonkka, G. A. Steinauer, K. J. Stoner
2022 Special Issue 1
57
sites. Transects were started at least 5 m (16.4 ft) from the property border to minimize
edge effects that may be caused by woodlands, roads, or other edge habitats frequently
associated with grassland boundaries. To avoid sampling in non-target habitats (e.g.,
woodlands, wetlands), it was sometimes necessary to adjust transect length and layout.
Following methods described in Oberhauser (2001) and using stems as a metric that represents
habitat restoration (MAFWA 2018), we counted milkweeds by their ramet (that is,
an individual stem of a clone). The two techniques we assessed were designed to require
similar field effort to directly compare the utility of the methods.
Belt-Transect Method. Transect locations for this sampling method were generated in
ArcMap 10.5. In order to sample 12–20% of the total accessible area, different widths of
transects were implemented depending on the site size. Site boundaries were delineated
according to land ownership because of access limitations. For larger areas, transects
were wider to maintain consistent percent coverage. The width of the transects ranged
from 1 to 4 m (3.28 to 13.12 ft). There were at least 10 transects per site with each
transect being at least 10 m (32.8 ft) apart. On larger sites (>40 ha), we sampled more
than 10 transects. A surveyor walked with a length of PVC pipe of the corresponding
measurement in order to visually demarcate the width of each transect. The surveyor
walked alertly between the start and end points of each transect, while using GPS points
as a guide. We recorded the total number of milkweed species and the number of stems
for each type.
Plot Method. For the plot method, we evenly placed points every 10 m (32.8 ft) along
the same transects that we had generated with ArcMap 10.5. We then used a random number
table to determine 100 of the plots to include in our sampling. For example, if the
number five was randomly generated, then we would sample the fifth point as the center
of a plot and continue in this manner. On site, we placed a 1-m2 quadrat of PVC pipe on
the ground to delineate the borders of our plot. Therefore, the area surveyed was 100 m2
(1,076 ft2) per site. The total number of milkweed stems was counted within each plot,
and stems were differentiated by species.
Statistical Analysis
For statistical analysis, we conducted negative binomial regression by using R Foundation
for Statistical Computing (R Core Team 2018) to model the over-dispersed milkweed
stem count data obtained by the plot sampling method and the belt-transect sampling
method in package MASS (Modern Applied Statistics with S; Venables and Ripley
2002). The response variable was the number of stems per hectare for each site. This was
calculated for each site by counting all stems in the belt-transect and dividing by the area
covered by the belt transect. For the plot method, it was calculated by counting all stems
in the plots and dividing by the area covered by all plots at the site. The observational
unit is the site for both methods, which is necessary to avoid pseudo-replication when
assessing differences in site type. Land-management type was included in the models
as a predictor. To test for differences in stem density among land-management types for
each sampling method, we used a likelihood ratio test to compare the full model to a
null model without the land-management type predictor. Post-hoc multiple comparisons
among land-management types were then calculated with a Tukey’s test and family-wise
error controlled by applying a Bonferroni correction. We then evaluated each regression
function (the regression function for stem counts obtained via plot sampling and that for
counts obtained via belt-transect sampling) to obtain predicted stem counts per hectare
for each land-management type, using parameter estimates from the models.
Prairie Naturalist
M. L. Manzanares, M. J. Panella, C. L. Wonkka, G. A. Steinauer, K. J. Stoner
2022 Special Issue 1
58
Land Management Type Milkweed Species Detected Milkweed Species Detected
via Plot Method via Belt-transect Method
High-diversity plantings AsSy, AsSu, AsVe, AsIn AsSy, AsSu, AsVe, AsIn, AsSp, AsViri
Pheasants Forever sites AsSy, AsVe AsSy, AsSu, AsVe, AsTu
Unmanaged pastures AsSy AsSy, AsSu, AsVe, AsIn, AsViri
Wildlife Management Areas AsSy, AsSu, AsVe, AsIn AsSy, AsSu, AsVe, AsIn, AsVir, AsTu
In addition, we tested for differences in the modeled stem counts obtained through plot
sampling versus those obtained through belt-transect sampling by combining the data from
both methods and using a negative binomial regression with both site type and sampling
type as predictors. We used a likelihood ratio test to compare this model to a reduced model
with only land-management type as a predictor to determine if modeled counts differed
between the two sampling methods. We also compared the milkweed richness detected by
plot and transect methods with a Poisson model; statistically, the richness data were not
over-dispersed. Richness (the number of milkweed species) was the response variable, and
sampling method was the predictor.
Results
Milkweed Summary
We surveyed a total of 48 properties and encountered over 25,000 milkweed stems across
the study. Species were comprised of Asclepias syriaca Linnaeus (Common Milkweed), A.
sullivantii Engelman ex A. Gray (Sullivant’s Milkweed), A. verticillata Linnaeus (Whorled
Milkweed), A. incarnata Linnaeus (Swamp Milkweed), A. speciosa Torrey (Showy Milkweed),
A. viridis Walter (Green antelopehorn), A. viridiflora Raf. (Green Comet Milkweed),
and A. tuberosa Linnaeus (Butterfly Weed) (Table 1). A. syriaca was the milkweed species
most often encountered, and A. tuberosa was least often encountered.
Transect Sampling Method
The negative binomial model was a good fit for milkweed stems per hectare obtained by
the belt-transect method. The expected 95% χ2 deviance on 44 degrees of freedom was 53.4,
and the model residual deviance was 56.3. The likelihood ratio test for the belt-transect
sampled model indicated that there was a difference among site types (χ2 = 56.3, P < 0.001).
Multiple comparisons showed that high-density planting sites had more milkweed stems
per hectare than wildlife management areas and the Pheasants Forever sites. We found that
unmanaged pastures had fewer milkweed stems per hectare than all other site types when
we used the transect method (Fig. 1, Table 2, Predicted counts Table 3). Predicted counts of
milkweed stems per hectare averaged across all site types was 324. Predicted richness was
2.08 milkweed species.
Plot Sampling Method
The negative binomial model was a good fit for milkweed stems per hectare obtained by the
plot method. The expected 95% χ2 deviance on 44 degrees of freedom was 53.4 and the model
Table 1. Species of milkweed detected on four different land management types by using two survey
methods in the eastern tallgrass prairie ecoregion of Nebraska. AsSy = Asclepias syriaca, AsSu = A.
sullivantii, AsVe = A. verticillata, AsIn = A. incarnata, AsSp = A. speciosa, AsViri = A. viridis, AsVir
= A. viridiflora, AsTu = A. tuberosa
Prairie Naturalist
M. L. Manzanares, M. J. Panella, C. L. Wonkka, G. A. Steinauer, K. J. Stoner
2022 Special Issue 1
59
residual deviances was 55.3. The likelihood ratio test for the plot-sampled model indicated that
there was a difference among site types (χ2 = 56.3, P < 0.001), and multiple comparisons revealed
that high-density planting sites and wildlife management areas had more milkweed stems
per hectare than did unmanaged pastures and the Pheasants Forever sites during our study (Fig.
2, Table 2, Predicted counts Table 3). Predicted counts of milkweed stems per hectare, averaged
across all site types, was 1675. Predicted richness was 1.08 milkweed species.
F i g u r e 1 . M i l k w e e d s t e m s
per hectare estimated using a
belt-transect sampling method
c o m p a r i n g d e n s i t i e s a c r o s s
four land-management types.
Different letters denote significant
differences among sites.
Figure 2. Stems of milkweed per
hectare estimated using a plot
samp l ing method compar ing
d e n s i t i e s a c r o s s f o u r l a n d -
management types. Different letters
denote significant differences
among sites.
Prairie Naturalist
M. L. Manzanares, M. J. Panella, C. L. Wonkka, G. A. Steinauer, K. J. Stoner
2022 Special Issue 1
60
Table 2. Bonferroni corrected multiple comparisons for plot stem counts and belt-transect stem counts
of species of Asclepias. PF= Pheasants Forever, HDP = high-diversity plantings, UP = unmanaged
pastures, and WMA = Wildlife Management Areas. The 0 in PF-HDP=0 indicates that our regression
was testing the null hypothesis that the number of stems in the PF sites minus the number of stems in
the HDP sites equals 0. The estimates are differences in log odds of expected counts among the site
types compared, reported here to two decimal places.
Linear Hypothesis Estimate Standard Deviation Z-score P-value
Table 3. Predictions of milkweed densities (stems per hectare) from two sampling methods
(i.e., plot and transect) for four different site types.
Plots
PF-HDP=0 -3.08 0.75 -4.12 <0.001
UP-HDP=0 -2.77 0.79 -3.52 0.003
WMA-HDP=0 -0.02 0.91 -0.02 1.0
UP-PF=0 0.31 0.83 0.38 1.0
WMA-PF=0 3.07 0.95 3.22 0.008
WMA-UP=0 2.75 0.98 2.80 0.30
Belt-transects
PF-HDP=0 -1.86 0.33 -5.64 <0.001
UP-HDP=0 -3.92 0.35 -11.00 <0.001
WMA-HDP=0 -1.31 0.40 -3.27 0.006
UP-PF=0 -2.06 0.37 -5.49 <0.001
WMA-PF=0 0.54 0.42 1.30 1.0
WMA-UP=0 2.60 0.44 5.93 <0.001
High-diversity plantings 3,194 736
Pheasants Forever sites 146 115
Unmanaged pastures 200 15
Wildlife management areas 3,143 198
Land Management Type Predictions for Plot Models
(milkweed stems per ha)
Predictions for Belt-transect
Models (milkweed stems per ha)
Sampling Method Comparison
We found differences between the two survey methods that we employed in their ability
to detect milkweeds. Across the 48 sites, we detected a sum of 804 milkweed stems when
using the plot method, whereas we detected a total well over 25,000 stems when using the
transect method. The count when using the transect method would of course have included
some of the same plants detected when using the plot method on the same sites on the same
day. While both plot and belt-transect models showed differences in milkweed stem density
across site types, the plot method appeared to be less effective at detecting differences
among sites than the belt-transect method. The negative binomial model was a good fit for
the combined plot and transect data, with sampling type as predictor for stems per hectare.
Prairie Naturalist
M. L. Manzanares, M. J. Panella, C. L. Wonkka, G. A. Steinauer, K. J. Stoner
2022 Special Issue 1
61
The expected 95% χ2 deviance on 91 degrees of freedom is 114.3, and the model residual
deviance was 118.0. The likelihood ratio test to compare models with and without sampling
type included as a predictor shows that milkweed stems per hectare differed by sampling
type (χ2 = 18.43, P < 0.001). The differences in log expected counts between transects and
plots was -1.6 ± 0.31 (Z = -5.1, P < 0.001). The numbers of predicted stems per hectare were
higher for plots than transects for all site types (Table 3).
The variance was much higher for stem counts obtained via plot sampling relative to
counts obtained via transects (Fig. 3). In fact, the standard deviation for the differences in
log odds of expected stem counts for plots was almost double that for transects (Table 2).
Milkweed richness was lower for plots than transects. The difference in log odds of milkweed
richness between transects and plots was 0.65 ± 0.17 (Z = 3.83, P < 0.001).
We did find similarities in the amount of survey effort expended using the belt-transect
method and the plot method. Each technique required a comparable amount of time and
energy to walk the site. On average, a site that was 40–60 ha would take about three hours
to survey 12–20% of a study area via transect or to sample 100 random plots.
Discussion
Detecting and sampling Asclepias species will likely remain an important objective to
describe host plant availability in the Monarch’s breeding habitat. Conservation programs
aim for a substantial increase in the number of milkweed stems (e.g., MAFWA 2018), while
grassland conversion continues to reduce available habitat (Thogmartin et al. 2017). The resulting
fluctuations in milkweed abundance necessitates periodic monitoring on lands under
diverse ownership and management to maintain a basic understanding of overall milkweed
availability and to make relevant recommendations for Monarch conservation actions.
As expected, we found differences in milkweed density and richness at sites according
to their management types. The high-diversity, local-ecotype plantings had the most species
of Asclepias, indicating survival of seeded milkweed and success for meeting Monarch
management objectives. The privately owned agricultural lands with grassland enhancement
(i.e., Nebraska Pheasants Forever land management) and the public wildlife management
area grassland rehabilitations (i.e., Nebraska Game and Parks Commission land management)
fared similarly in terms of species of milkweed per ha as indicated under belt-transect
F i g u r e 3 . C omp a r i s o n o f
milkweed densities (milkweed
stems per hectare) by site type
using the plot method and the
belt-transect method.
Prairie Naturalist
M. L. Manzanares, M. J. Panella, C. L. Wonkka, G. A. Steinauer, K. J. Stoner
2022 Special Issue 1
62
sampling. Richness was higher in all seeded site types, and milkweed species that were
not part of the original seeding mixtures were even occasionally detected at the previously
planted sites. Unsurprisingly, pastures that were not managed specifically for milkweed, or
for pollinator resources in general, had lower diversity and abundance of milkweeds.
Our results suggest that by choosing a diverse seed mix of native plants with a milkweed
component and undertaking conservation-based management practices, a land manager
increases the chances of providing Monarch breeding habitat. Similarly, Zaya et al.
(2017) reported that agricultural areas that were not managed for pollinators had fewer
milkweeds than natural areas (e.g., grasslands, wetlands, and forests) in Illinois. They also
found that sites planted with low-diversity seed mixes or those frequently mowed or sprayed
with herbicides, as well as those that are overgrazed, are poor Monarch habitat. They promoted
management practices that support plant diversity at various successional states to
improve habitat for A. syriaca, In Iowa, Kaul and Wilsey (2019) found higher densities of
milkweed species in remnant native prairies that had higher forb diversity. In our study area,
pastures that were not managed for pollinator benefits had fewer milkweeds, and thus cannot
be expected to support much Monarch breeding without some degree of management.
This finding is likely associated with grazing regimes and other land management practices
that may reduce native plant diversity, and results could differ in other locations as these
variables change.
Monitoring efforts in the Midwest to count milkweeds and assess Monarch breeding
habitat have employed a variety of methods, highlighting the myriad options available to
managers. For example, the Integrated Monarch Monitoring Protocol (IMMP) and Monarch
Larva Monitoring Project (MLMP) (Monarch Joint Venture 2021a, 2021b) include wellknown
community science programs. Lukens et al. (2020) used the IMMP in Minnesota,
Wisconsin, and Iowa. Kasten et al. (2016) used a modified interrupted belt-transect sampling
method where they randomly generated starting points for milkweed roadside surveys
within a 402.34 m (250 mi) radius of Minneapolis, Minnesota. Kaul and Wilsey (2019)
chose a randomly located 3.14-m2 round plot method in Iowa. Thogmartin et al. (2017)
used spatial information from the 2014 Cropland Data Layer (CDL; USDA NASS 2021)
and other sources in order to develop a land-cover map that was used to estimate milkweed
density in all or portions of several mid-western states. The US Geological Survey Upper
Midwest Environmental Sciences Center (UMESC 2021) considers multiple methods to
meet the challenges of Monarch conservation.
With limited resources, conservation program managers want to select, and surveyors
want to use, the most appropriate and effective method for their circumstances. Individuals
charged with the task of milkweed monitoring must take into consideration the pros and
cons of their survey choices. Land managers voice common concerns regarding the amount
of time, training, human resources, and funding required to conduct milkweed counts on
a variety of land management types. They also express challenges related to working during
the peak of summer (e.g., hotter temperatures, higher ultraviolet radiation exposure),
while maximizing outcomes to get accurate measurements during periods of peak milkweed
blooming to aid in its detection and identification. Assessing the different methods available
as a pilot to the development of a monitoring approach can help maximize efficiency in the
face of these concerns.
We assessed only two methods on four different types of properties in the tallgrass
prairies of eastern Nebraska, and it was clear that the methods did not perform equally.
Our results indicated that the belt-transect technique allowed us to survey more area and
exhibited lower variation; therefore, it seems more useful in describing species abundance
Prairie Naturalist
M. L. Manzanares, M. J. Panella, C. L. Wonkka, G. A. Steinauer, K. J. Stoner
2022 Special Issue 1
63
and richness at our sites than the plot sampling method. We were also able to implement
the belt-transect method with a similar amount of time and effort in the field, compared
to the plot method. Thus, it was not costlier or more difficult to implement while at the
site, although a bit more time was needed in making planning preparations to visit a particular
site. A biologist needed to review aerial imagery of the survey sites to first find the
total area of a site, and then digitally plan placement of the transect coordinates in order
to effectively survey at least 12% of a total site. As this was part of the biologist’s other
concurrent responsibilities, we did not track the time or cost per hour associated with
this milkweed survey planning activity as it was being completed. This would make for
an interesting analysis for future work to further compare the relative benefits of the two
types of survey methods.
Perhaps, our findings also reflect the biological characteristics of Asclepias species.
Milkweed seeds are often dispersed by wind, but several milkweed species can reproduce
asexually by vegetative cloning and spreading rhizomes. For example, Common Milkweed
and Showy Milkweed are rhizomatous species (Luna and Dumroese 2013) that allow
for spreading growth in a “clumpy” (i.e., patchy) manner. Likewise, Kasten et al. (2016)
noted the patchy distribution of milkweeds across the landscape during their multi-state
roadside surveys. This tendency for clumping growth habit is not dependent on seed dispersal
by wind and could help explain why the belt-transect method was more effective
than the plot method at detection of milkweeds at our sites. Therefore, managers should
also consider the biology of the milkweed species likely to be found in their region in
selecting monitoring methods.
Literature Cited
Borders, B., and E. Lee-Mäder. 2014. Milkweeds: A conservation practitioner’s guide. The Xerces Society
for Invertebrate Conservation. Portland, Oregon, USA. Available online at https://xerces.org/sites/
default/files/2018-05/17-031_02_XercesSoc_Milkweeds-Conservation-Guide_web.pdf. Accessed 2
Aug 2021.
Brower, L.P. 1984. Chemical defense in butterflies. Pages 109–134 In R.I. Vane Wright and P.R. Ackery,
(Eds.). The Biology of Butterflies. Symposium of the Royal Entomological Society of London 11.
Academic Press London, England.
Brower, L.P., O.R. Taylor, E.H. Williams, D.A. Slayback, R.R. Zubieta, and M.I. Ramírez. 2012. Decline
of monarch butterflies overwintering in Mexico: Is the migratory phenomenon at risk? Insect Conservation
and Diversity 5:95–100.
Flockhart, D.T., J. Pichancourt, D.R. Norris, and T.G. Martin. 2014. Unravelling the annual cycle in a
migratory animal: Breeding-season habitat loss drives population declines of monarch butterflies.
Journal of Animal Ecology 84:155–165.
Hartzler, R.G., and D.D. Buhler. 2000. Occurrence of common milkweed (Asclepias syriaca) in cropland
and adjacent areas. Crop Protection 19:363–366.
Kasten, K., C. Stenoien, W. Caldwell, and K.S. Oberhauser. 2016. Can roadside habitat lead monarchs
on a route to recovery? Journal of Insect Conservation 20:1047–1057.
Kaul, A.D., and B.J. Wilsey. 2019. Monarch butterfly host plant (milkweed Asclepias spp.) abundance
varies by habitat type across 98 prairies. Restoration Ecology 27:1274–1281.
Lukens, L., K. Kasten, C. Stenoien, A. Cariveau, W. Caldwell, and K. Oberhauser. 2020. Monarch Habitat
in Conservation Grasslands. Frontiers in Ecology and Evolution. 8(13)1–13.
Luna, T., and R.K. Dumroese. 2013. Monarchs (Danaus plexippus) and milkweeds (Asclepias species):
The current situation and methods for propagating milkweeds. Native Plants Journal 14:5–15.
Midwest Association of Fish and Wildlife Agencies (MAFWA). 2018. Mid-America monarch conservation
strategy 2018–2038. Version 1.0. Available online at www.mafwa.org/wp-content/uploads/
2018/07/MAMCS_June2018_Final.pdf. Accessed 6 Jul 2021.
Prairie Naturalist
M. L. Manzanares, M. J. Panella, C. L. Wonkka, G. A. Steinauer, K. J. Stoner
2022 Special Issue 1
64
Monarch Joint Venture. 2021a. Integrated Monarch Monitoring Program (IMMP). Monarch Joint Venture,
St. Paul, Minnesota, USA. Available online at https://monarchjointventure.org/images/uploads/documents/
IMMP_Guidebook_2021.pdf. Accessed 23 Jul 2021.
Monarch Joint Venture. 2021b. Monarch Larva Monitoring Project (MLMP). Monarch Joint Venture, St.
Paul, Minnesota, USA. Available online at https://monarchjointventure.org/our-work/monarch-larvamonitoring-
project-online-training. Accessed 26 Jul 2021.
Nabhan, G.P., I. Warren, and O.R. Taylor. 2015. Monarch recovery from a milkweed’s point of view.
Make way for monarchs, Brevard, North Carolina, USA. Available online at https://makewayformonarchs.
org/2015/10/monarch-recovery-from-a-milkweeds-point-of-view. Accessed 2 Aug 2021.
Oberhauser, K. S., M. D. Prysby, H. R. Mattila, D. E. Stanley-Horn, M. K. Sears, G. Dively, and R. L.
Hellmich. 2001. Temporal and spatial overlap between monarch larvae and corn pollen. Proceedings
of the National Academy of Sciences 98:11913–11918.
Panella, M.J. (Ed.). 2017. Conservation strategy for Monarchs (Danaus plexippus) and at-risk pollinators
in Nebraska. Nebraska Monarch and Pollinator Initiative, Nebraska Game and Parks Commission,
Lincoln, Nebraska, USA. Available online at http://outdoornebraska.gov/wp-content/uploads/2018/04/
Monarch-Pollinator-Plan_NE_Sept2017-1.pdf. Accessed 6 Jul 2021.
Pleasants, J. 2017. Milkweed restoration in the Midwest for monarch butterfly recovery: Estimates of
milkweeds lost, milkweeds remaining and milkweeds that must be added to increase the monarch
population. Insect Conservation and Diversity 10:42–53.
Pleasants, J.M., and K.S. Oberhauser. 2012. Milkweed loss in agricultural fields because of herbicide use:
Effect on the monarch butterfly population. Insect Conservation and Diversity 6:135–144.
R Core Team. 2018. R: A language and environment for statistical computing. R Foundation for Statistical
Computing, Vienna, Austria. Available online at https://www.r-project.org/. Accessed 2 August 2021.
Schneider, R., K. Stoner, G. Steinauer, M. Panella, and M. Humpert (Eds.). 2011. The Nebraska Natural
Legacy Project: State Wildlife Action Plan. (2nd Edition) The Nebraska Game and Parks Commission,
Lincoln, Nebraska, USA. Available online at http://outdoornebraska.gov/wp-content/uploads/2015/09/
NebraskaNaturalLegacyProject2ndEdition.pdf. Accessed 2 Aug 2021.
Semmens, B.X., D.J. Semmens, W.E. Thogmartin, R. Wiederholt, L. López-Hoffman, J.E. Diffendorfer,
and O.R. Taylor. 2016. Quasi-extinction risk and population targets for the eastern, migratory population
of monarch butterflies (Danaus plexippus). Scientific Reports 6(1). doi:10.1038/srep23265.
Stenoien, C., K.R. Nail, J.M. Zalucki, H. Parry, K.S. Oberhauser, and M.P. Zalucki. 2016. Monarchs in
decline: A collateral landscape-level effect of modern agriculture. Insect Science 25:528–541.
Thogmartin, W.E., L. López-Hoffman, J. Rohweder, J. Diffendorfer, R. Drum, D. Semmens, S. Black,
I. Caldwell, D. Cotter, P. Drobney, L.L. Jackson, M. Gale, D. Helmers, S. Hilburger, E. Howard, K.
Oberhauser, J. Pleasants, B. Semmens, O. Taylor, P. Ward, J.F. Weltzin, and R. Wiederholt. 2017.
Restoring monarch butterfly habitat in the Midwestern US: “All hands on deck”. Environmental
Research Letters 12:074005. doi: 10.1088/1748-9326/aa,7637.
US Department of Agriculture Farm Service Agency (USDA FSA) 2021. Common Land Unit (CLU).
Available online at https://www.fsa.usda.gov/programs-and-services/aerial-photography/imageryproducts/
common-land-unit-clu/index. Accessed 20 December 2021.
US Department of Agriculture National Agricultural Statistics Service (USDA NASS). 2021. Cropland
Data Layer. Available online at https://www.nass.usda.gov/Research_and_Science/Cropland/SARS1a.
php. Accessed 21 December 2021.
US Geological Survey Upper Midwest Environmental Sciences Center (UMESC). 2021. Monarch
Conservation Science Partnership. Available online at https://www.usgs.gov/centers/umesc/science/
monarch-conservation-science-partnership?qt-science_center_objects=0#qt-science_center_objects.
Accessed 26 Jul 2021.
Venables, W.N., and B.D. Ripley. 2002. Modern Applied Statistics with S. Fourth Edition. Springer, New
York, USA. 495 pages.
Vidal O., and E. Rendón-Salinas. 2014. Dynamics and trends of overwintering colonies of the monarch
butterfly in Mexico. Biological Conservation 180:165–175.
Zaya, D.N., I.S. Pearse, and G. Spyreas. 2017. Long-term trends in midwestern milkweed abundances and
their relevance to monarch butterfly declines. BioScience 67:343–356.