Small Mammal Assemblages in Natural Plant Communities of Wisconsin
File(s)
Date
2012-06Author
Stephens, Ryan B.
Publisher
University of Wisconsin-Stevens Point, College of Natural Resources
Metadata
Show full item recordAbstract
Small mammals play important roles in Wisconsin ecosystems; however, considering
their importance and ubiquity, we know little about the best way to capture them or how to
distinguish between cryptic species. Additionally, we do not have consistent information
regarding distributions and habitat requirement of many small mammal species. This study
was broken into 4 parts in order to address these knowledge gaps.
When evaluating richness and diversity of small mammal communities it is
important to consider the impact that trap efficacy may have on these indices. The
objectives of my study were to determine species specific trap efficacy relative to
Sherman traps and pitfall traps, assess the impact of trap efficacy on measures of species
richness and diversity, and compare mortality rates between trap types, and if pitfall
covers reduce trap mortality. In the summers of 2009 and 2010, I trapped throughout
Wisconsin in 5 habitats. I used 180 transects (190 m-long) of 20 Sherman live traps
spaced every 10 m and 10 pitfall traps spaced every 20 m for 4 consecutive nights. I
trapped 3,261 small mammals of 22 species in 34,235 trap nights. Pitfall traps were more
effective at capturing shrews and voles, whereas Sherman traps captured more mice
(Peromyscus spp.) and squirrels. Irrespective of habitat type, both trap types together
captured significantly higher species richness and diversity than either trap type
independently. Covers significantly reduced mortality for Peromyscus spp., but not for
voles or shrews and covers reduced overall captures of voles. My results indicate that
Sherman and pitfall traps capture different portions of the small mammal community and,
regardless of the habitat type, should be used in combination when assessing species
richness and diversity.
In Wisconsin, white-footed mice (Peromyscus leucopus noveboracensis) and
woodland deer mice (P. maniculatus gracilis) are difficult to distinguish. Recent climatic
trends have facilitated encroachment of P. leucopus north into the range of P.
maniculatus, making unambiguous species identification imperative. Cranial and
external measurements have been used previously to differentiate these species.
However, since large geographic morphological variation occurs and most previous
studies used measurements from deceased specimens, definitive morphological
measurements need to be identified that can quickly and effectively classify live
Wisconsin Peromyscus spp. in the field. I live trapped Peromyscus spp. in central and
northern Wisconsin during the summer of 2010. I measured ear, tail, hindfoot, and
weight from live animals and collected tissue for genetic confirmation. In addition to
using mDNA analysis to identify samples to species, I collected measurements from 84 P.
maniculatus and 293 P. leucopus in 6 Wisconsin counties. I used discriminate function
analysis (DFA) to develop equations to identify characteristics that best separated mice to
species. By using ear length alone, 97.9% of the samples could be correctly classified
with most P. leucopus <17mm and most P. maniculatus ≥ 17mm. By adding weight to
the function, I was able to achieve 99.2% classification accuracy and with the addition of
tail was able to achieve 99.5% differentiation.
The arctic shrew (Sorex arcticus) and American water shrew (S. palustris) are
wide-spread boreal species. Sorex arcticus reaches its southern distribution in Wisconsin
and within the Midwest S. palustris also reaches its southern range extent in Wisconsin.
Although the ranges of both species are generally thought to occur in the Driftless Area
in the southwestern portion of Wisconsin, no occurrences have been documented from
the region. Within the past decade and half, the known distribution of these species has
changed dramatically. I used museum records and geo-referenced surveys to remap the
known distribution of S. arcticus and S. palustris in Wisconsin. Herein, I report records
of S. arcticus and S. palustris within the Driftless Area and extend the distribution of S.
arcticus and S. palustris south of their former range, including the most southerly extant
record of S. arcticus. This study emphasizes the continued value of voucher specimens
and museum collections.
Determining small mammal species habitat associations and environmental
characteristics, important in site occupancy, are central to understanding species biology
and community organization. During the summers of 2009 and 2010, I trapped small
mammals and measured habitat variables throughout the state of Wisconsin at 180 sites
spanning 13 natural habitats. I captured 3,261 individual small mammals of 22 species,
allowing for characterization of small mammal communities within natural habitats. I
was also able to model occupancy for 12 species using habitat covariates while
incorporating imperfect detection. Comparisons among small mammal communities
indicated that organization was based on soil moisture, abundance of trees, and climate.
Additionally, the most important covariates in predicting small mammal occupancy were
basal area, frequency and duration of wet periods, and minimum winter temperature. The
ability to detect differences in community similarity and species occupancy, based on
climate, is likely a product of the large geographical extent of this study, underscoring the
need to assess community structure at multiple scales. Given the importance of climate
in structuring these small mammal communities, predicted warming winter temperatures
could have considerable impacts on small mammal communities in the region.
Interestingly, occupancy of the meadow jumping mouse was best predicted by the
presence of precipitation before a trap check and availability of pitfall traps. These
conclusions would have been missed using other analyses that do not account for
imperfect detection. Finally, my results indicate that soil moisture, abundance of trees,
and climate are important for species occupancy and, at the community level, these
variables create regular and predictable community structure across a heterogeneous
landscape.