Remnant Oak Savanna Research Articles

Avian Nest Success along a Habitat Gradient in the Cross Timbers Oak Savanna

Abstract Oak savanna, once widespread across central North America, has functionally vanished from most of its range due to land conversion or fire suppression and subsequent afforestation. Savanna-associated bird species have exhibited population declines as a result. One of the few areas containing remnant oak savanna is the vast Cross Timbers ecoregion of the south-central Unites States. Previous research in the Midwest showed some bird species exhibited higher rates of nest success (i.e., lower predation of eggs and nestlings) in restored savannas than in closed-canopy forests. Our objective was to quantify patterns of avian nest success across a gradient from open-canopy oak savanna to closed-canopy forest in the Cross Timbers ecoregion of southeastern Kansas. Daily nest survival rates for four common bird species—mourning dove (Zenaida macroura), yellow-billed cuckoo (Coccyzus americanus), brown thrasher (Toxostoma rufum), and northern mockingbird (Mimus polyglottos)—were modeled against covariates ...

Testing the use of best professional judgment to create biological benchmarks for habitat assessment of wetlands and oak savannas in northwestern Indiana

The Tolleston Strandplain at the southern end of Lake Michigan offers a unique “dune and swale” topography supporting oak savannas on the dunes and a mosaic of wetland communities in the swales. Following years of human degradation, sites in this area are now being restored. In this effort, assessments of vegetative quality have been necessary for proper management decisions. However, it is poorly understood what indices best reflect the vegetative quality of these oak savannas and wetlands. A potential method for determining the best indices for these community types is to use metric benchmarks that employ expert best professional judgment (BPJ). In order to confirm the viability and consistency of BPJ for creating benchmarks, Kappa analysis was used to determine the level of agreement among seven experts. They placed each of 63 transects from this unique landscape into one of four quality categories: (1) “good to very good,” (2) “medium,” (3) “poor,” and (4) “very poor.” While experts had good agreement about the quality of severely degraded riverine wetlands, they had fair agreement when assessing the swales, and poor agreement when assessing the oak savannas. Using discriminant analysis, follow-up questions, and a comparison of each expert's quality categories with remnant oak savanna metrics, varying perspectives of quality also were discovered which may have influenced the experts’ assessments of the sites. Therefore, BPJ must be used with caution when creating metric benchmarks.

Seedling Recruitment and Establishment ofLupinus perennisin a Mixed-Management Landscape

Ecological disturbances structure plant communities. In areas where reintroducing natural disturbances is unfeasible, land managers use artificial disturbances, such as fire and mowing, to simulate natural disturbances. However, prescribed fire and mowing can promote or inhibit seedling survival in plant populations. Lupinus perennis (the wild blue lupine) is a potentially threatened species commonly subjected to rotations of burning and mowing treatments to restore habitat for endangered butterflies (e.g., Karner blue, Lycaeides melissa sameulis). While L. perennis adults respond favorably to these management techniques, it is unknown how these variable management programs affect natural seedling recruitment and establishment. We quantified natural L. perennis seedling survival in seven managed remnant oak savannas in Northwest Ohio in 2007 and 2008. To evaluate the roles of management history and environmental factors on recruitment, we obtained burning and mowing management records for each savanna, measured environmental variables at each seedling, and assessed their relationships with seedling mortality. Seedling survival varied among the savannas and increased with greater fern and grass stems, moss cover, soil moisture, and initial size, but was reduced by increased oak saplings, light levels, and topographic wetness. Site elevation and initial seedling size affected the sizes of the juvenile plants. Management activities did not directly influence seedling survival or density, but the number of fires and total management activities at a site were positively correlated with oak saplings, suggesting current management practices may indirectly reduce lupine survival. Future management strategies may increase recruitment by focusing on the restoration of multiple species that are important components of the lupine regeneration niche.

Negative native–exotic diversity relationship in oak savannas explained by human influence and climate

Recent research has proposed a scale‐dependence to relationships between native diversity and exotic invasions. At fine spatial scales, native–exotic richness relationships should be negative as higher native richness confers resistance to invasion. At broad scales, relationships should be positive if natives and exotics respond similarly to extrinsic factors. Yet few studies have examined both native and exotic richness patterns across gradients of human influence, where impacts could affect native and exotic species differently. We examined native–exotic richness relationships and extrinsic drivers of plant species richness and distributions across an urban development gradient in remnant oak savanna patches. In sharp contrast to most reported results, we found a negative relationship at the regional scale, and no relationship at the local scale. The negative regional‐scale relationship was best explained by extrinsic factors, surrounding road density and climate, affecting natives and exotics in opposite ways, rather than a direct effect of native on exotic richness, or vice versa. Models of individual species distributions also support the result that road density and climate have largely opposite effects on native and exotic species, although simple life history traits (life form, dispersal mode) do not predict which habitat characteristics are important for particular species. Roads likely influence distributions and species richness by increasing both exotic propagule pressure and disturbance to native species. Climate may partially explain the negative relationship due to differing climatic preferences within the native and exotic species pools. As gradients of human influence are increasingly common, negative broad‐scale native–exotic richness relationships may be frequent in such landscapes.

Patterns in bird community structure related to restoration of Minnesota dry oak savannas and across a prairie to oak woodland ecological gradient

There is limited understanding of the influence of fire and vegetation structure on bird communities in dry oak (Quercus spp.) savannas of the Upper Midwest and whether bird communities in restored savanna habitats are similar to those in remnant savannas. During the 2001 and 2002 breeding seasons, we examined the relationship between bird communities and environmental variables, including vegetation characteristics and site prescribed-burn frequencies, across a habitat gradient in dry oak savannas in central Minnesota. The habitat gradient we studied went from: (1) prairie to (2) remnant oak savanna to (3) oak woodland undergoing savanna restoration via fire or mechanical removal of woody vegetation to (4) oak woodland. We conducted fixed-radius point counts (n = 120) within habitats with either prairie groundcover or predominately oak canopy. We described canopy and groundcover characteristics at a sub-sample (n = 28) of non-prairie points, and collected canopy and woody species richness data and prescribed-burn frequencies over the past 20 years for all points. Observed bird communities were most strongly correlated with canopy cover and burn frequency and, to a lesser extent, attributes of the shrub component. Most savanna points had bird communities that were distinct from those found at oak woodland or oak woodland points undergoing restoration via burning. Savanna points similar to oak woodland points were in areas managed by periodic cutting rather than burning. Remnant savanna bird communities were more strongly associated with prescribed burning than those in other habitat types, but it appeared that most oak woodlands that had undergone ≥ 20 years of prescribed burning remained ecologically distinct from remnant savannas. This suggests that some savannas that have converted to oak woodlands may exist in an alternative, or stable, ecological state even following extended restoration efforts.

Environmentally biased fragmentation of oak savanna habitat on southeastern Vancouver Island, Canada

Quantifying the degree to which natural or protected areas are representative of a specified baseline provides critical information to conservation prioritization schemes. We report results on southeastern Vancouver Island, Canada, where we compared environmental conditions represented across the entire landscape, in oak savanna habitats prior to European settlement (<1850), and in both protected and unprotected oak savannas in the present-day. In this region, oak savannas represent a rare habitat type, harboring many threatened species. Before European settlement, oak savannas occurred in a distinctly different subset of environmental conditions than they do today. Compared to the entire landscape, oak savannas were historically found predominantly in warm, dry, flat, and low-lying areas, but habitat destruction has left oak savannas in largely the exact opposite set of conditions at present. Thus, the range of conditions in both protected and unprotected oak savannas at present are highly unrepresentative of historical conditions. It appears that fire management by indigenous peoples maintained oak savannas historically across large areas of flat low-lying conditions with deep soils, where succession otherwise produces closed coniferous forest. These areas have since been almost entirely converted to agricultural and urban areas, leaving remnant oak savannas largely on steep, rocky hilltops, where the habitat is maintained by shallow soils. Our results provide quantitative guidance for setting conservation priorities for oak savannas in this region, while highlighting the important general issue of the major role traditional land-use practices can play in shaping landscapes, and therefore in influencing the baselines used to set conservation priorities.

Effects of Disturbance Frequency on Canopy Age Structure and Species Composition in a Managed Midwestern Oak Savanna

In the Midwest, the oak savanna is a unique plant community located between the western prairies and northern forests, but, once the most common plant community in Wisconsin, it is now one of the region's rarest plant communities. To assess the success of varying intensities of management practices, standard dendrochronology and vegetation surveying methods were used to determine the age structure and species composition in oak savanna sites that varied from intensely managed to unmanaged. Size and age distributions revealed that the number of distinct cohorts of oaks increased in less managed sites. Seedlings and saplings were present in high frequencies in all sites; however, management practices such as fire applications that were too frequent apparently decreased oak survivorship, as did the competition in the closed canopies of the unmanaged sites. A fire frequency of 2-3 years eliminates competition, but a frequency of 4-6 years, typical of fire-adapted forest communities, allows for slower-growing species and younger oaks to survive, while still maintaining a low density. In these conditions, where growth rates vary in time, the use of dendrochronology is necessary to obtain an accurate canopy structure for fire-adapted forest communities. Quercus macrocarpa was dominant in the more managed sites, but shade-tolerant species were dominant in the unmanaged site. At present, the canopy composition in the managed sites represented an oak savanna assemblage; however, the unmanaged composition was shifting to a woodland assemblage. Thus, remnant oak savanna can be over- or under-managed, and unmanaged oak savanna can be overlooked as woodlands.

A 200-year Fire History in a Remnant Oak Savanna in Southeastern Wisconsin

Abstract In oak (Quercus) savannas in the Midwest region of the United States, fire continues to play a significant role in its persistence within the landscape; however, in southeastern Wisconsin, quantified fire history records are limited. This study documented occurrences of fire for the last 200 y using 16 cross-dated oak sections from an oak savanna remnant in Kenosha County, Wisconsin. Fire history was reconstructed by: (1) obtaining fire-scarred cross sections from cut down oaks in an area planned for an industrial park, (2) preparing sample surfaces to count rings, (3) creating skeleton plots to crossdate samples, (4) dating fire scars, (5) calculating intervals between fires and (6) determining seasonality of the fire. In addition, a master chronology was constructed for the oak savanna using the cross sections to correlate with precipitation. Throughout the entire time series, only 50% of the fire dates were associated with significantly dry years. Although the overall Weibull median interval w...

Scale‐dependence of environmental effects on species richness in oak savannas

Abstract. In order to investigate how scale (grain size) affects the relationships between species richness and environmental drivers (such as stress and disturbance), we collected 12 nested quadrats (from 0.25 m2 to 1023 m2) from seven remnant oak savannas located in the floodplain of the Chippewa River in western Wisconsin, USA. Large and small‐scale richness were not significantly correlated, suggesting that small‐scale richness is not strongly controlled by sampling effects of the local species pool. Linear and curvilinear regressions between species richness and disturbance, canopy cover, biomass, and soil organic matter were dependent on sampling scale (grain size). Disturbance by fire was strongly related to richness at small scales, while tree canopy cover was strongly related to richness at larger scales. While there was some evidence suggesting the transition from disturbance to canopy effects occurs between 10 and 100 m2, the transition was not particularly abrupt. The results cast doubt on the general importance of local species pools in affecting small‐scale richness as well as our ability to make generalizations that do not explicity include scale.

Scale-dependence of environmental effects on species richness in oak savannas

Abstract. In order to investigate how scale (grain size) affects the relationships between species richness and environmental drivers (such as stress and disturbance), we collected 12 nested quadrats (from 0.25 m2 to 1023 m2) from seven remnant oak savannas located in the floodplain of the Chippewa River in western Wisconsin, USA. Large and small-scale richness were not significantly correlated, suggesting that small-scale richness is not strongly controlled by sampling effects of the local species pool. Linear and curvilinear regressions between species richness and disturbance, canopy cover, biomass, and soil organic matter were dependent on sampling scale (grain size). Disturbance by fire was strongly related to richness at small scales, while tree canopy cover was strongly related to richness at larger scales. While there was some evidence suggesting the transition from disturbance to canopy effects occurs between 10 and 100 m2, the transition was not particularly abrupt. The results cast doubt on the general importance of local species pools in affecting small-scale richness as well as our ability to make generalizations that do not explicity include scale.

Gradients in the Composition, Structure, and Diversity of Remnant Oak Savannas in Southern Wisconsin

Savannas with scattered oak canopies, once the most widespread communities in southern Wisconsin, now are among the most endangered. Surprisingly little is known about the composition, structure, and horizontal patterning of their species-rich ground layers. This study relates the distribution and ecological characteristics of 417 ground-layer species to local and regional gradients in soil composition and light regime, based on an analysis of 722 1-m2 quadrats in 12 remnant savannas. Our findings have important implications for efforts to restore/conserve midwestern oak savannas. Ground-layer composition was strongly related to among-site differences in soil texture and within-site differences in light availability, with variation in sand vs. silt content accounting for twice as much turnover in species composition as that accounted for by direct photon flux density (PFD) estimated from hemispherical photographs. Most species reached peak coverage under sunny or partly shaded conditions. Flowering/fruiting was often skewed toward sunnier microsites. Absolute forb cover increased with silt content and declined with PFD. Graminoid cover showed a curvilinear relationship to soil texture and light, being highest in well-lit, moderately sandy microsites. Total ground-layer cover increased with silt content at a given irradiance; it increased with PFD on silty sites and decreased with PFD on sandy sites. Forb cover increased regularly with PFD and sand content. When quadrats were stratified by sand content and PFD, species richness of forbs and graminoids increased linearly with coverage by each group, with far more forb species present at a given coverage. Among graminoids, C4 grasses were common only in bright, sandy microsites; C3 grasses and sedges had broader ecological distributions. Among forbs, leaf width increased and leaf inclination became more horizontal toward shadier and siltier microsites; tall herbs were generally found in silty areas with dense ground-layer cover. Plants with N-fixing symbioses were found mostly on sandy, well-lit microsites, although climbing species occured on shadier and/or siltier microsites, where N-fixation might be expected to be less advantageous. Most ground-layer species were perennial (88%), with few annuals (6%) or biennials (6%). For the 85 most common species, breadth of distribution across savanna microhabitats (4 soil × 4 light) was significantly correlated with presence across 34 Wisconsin community types, suggesting that similar factors help to constrain distributions at local and regional scales. Species in the two largest genera (Aster and Solidago) differed significantly in distribution according to the Syrjala test, supporting habitat partitioning as a mechanism of coexistence. Oak savannas are unusually diverse. At small spatial scales, there were 16.1 ± 1.3 species/m2 (mean ± 1 sd), compared with 11.4 ± 2.7 for prairies and 8.2 ± 2.5 for forests. At slightly larger scales, savannas showed high spatial turnover in ground-layer composition: 89.3 ± 12.0 species/20 m2, compared with 41.5 ± 8.3 for prairies and 42.8 ± 8.0 for forests. At large spatial scales, a survey of 22 savanna remnants (42 ha) showed extensive floristic differentiation: 507 native plant species, ∼27% of Wisconsin's indigenous vascular flora (∼14 × 106 ha). Contrary to previous reports, Midwest oak savannas are forb dominated, except on the sandiest or sunniest microsites. Release of forbs from competition with C4 grasses under partly shaded conditions may help to account for the high diversity of savanna ground layers relative to prairies. Divergent distributions of plants with different adaptations for energy capture, together with large variation within sites in ground-layer light regime and among sites in soil texture, suggest that partitioning of light and soil gradients is important for maintaining the high plant diversity of oak savannas. Mass effects (involving dispersal subsidies from favorable microsites) may also play a role in maintaining diversity within these mosaic communities.

GRADIENTS IN THE COMPOSITION, STRUCTURE, AND DIVERSITY OF REMNANT OAK SAVANNAS IN SOUTHERN WISCONSIN

Savannas with scattered oak canopies, once the most widespread communities in southern Wisconsin, now are among the most endangered. Surprisingly little is known about the composition, structure, and horizontal patterning of their species-rich ground layers. This study relates the distribution and ecological characteristics of 417 ground-layer species to local and regional gradients in soil composition and light regime, based on an analysis of 722 1-m2 quadrats in 12 remnant savannas. Our findings have important implications for efforts to restore/conserve midwestern oak savannas. Ground-layer composition was strongly related to among-site differences in soil texture and within-site differences in light availability, with variation in sand vs. silt content accounting for twice as much turnover in species composition as that accounted for by direct photon flux density (PFD) estimated from hemispherical photographs. Most species reached peak coverage under sunny or partly shaded conditions. Flowering/fruiting was often skewed toward sunnier microsites. Absolute forb cover increased with silt content and declined with PFD. Graminoid cover showed a curvilinear relationship to soil texture and light, being highest in well-lit, moderately sandy microsites. Total ground-layer cover increased with silt content at a given irradiance; it increased with PFD on silty sites and decreased with PFD on sandy sites. Forb cover increased regularly with PFD and sand content. When quadrats were stratified by sand content and PFD, species richness of forbs and graminoids increased linearly with coverage by each group, with far more forb species present at a given coverage. Among graminoids, C4 grasses were common only in bright, sandy microsites; C3 grasses and sedges had broader ecological distributions. Among forbs, leaf width increased and leaf inclination became more horizontal toward shadier and siltier microsites; tall herbs were generally found in silty areas with dense ground-layer cover. Plants with N-fixing symbioses were found mostly on sandy, well-lit microsites, although climbing species occured on shadier and/or siltier microsites, where N-fixation might be expected to be less advantageous. Most ground-layer species were perennial (88%), with few annuals (6%) or biennials (6%). For the 85 most common species, breadth of distribution across savanna microhabitats (4 soil × 4 light) was significantly correlated with presence across 34 Wisconsin community types, suggesting that similar factors help to constrain distributions at local and regional scales. Species in the two largest genera (Aster and Solidago) differed significantly in distribution according to the Syrjala test, supporting habitat partitioning as a mechanism of coexistence. Oak savannas are unusually diverse. At small spatial scales, there were 16.1 ± 1.3 species/m2 (mean ± 1 sd), compared with 11.4 ± 2.7 for prairies and 8.2 ± 2.5 for forests. At slightly larger scales, savannas showed high spatial turnover in ground-layer composition: 89.3 ± 12.0 species/20 m2, compared with 41.5 ± 8.3 for prairies and 42.8 ± 8.0 for forests. At large spatial scales, a survey of 22 savanna remnants (42 ha) showed extensive floristic differentiation: 507 native plant species, ∼27% of Wisconsin's indigenous vascular flora (∼14 × 106 ha). Contrary to previous reports, Midwest oak savannas are forb dominated, except on the sandiest or sunniest microsites. Release of forbs from competition with C4 grasses under partly shaded conditions may help to account for the high diversity of savanna ground layers relative to prairies. Divergent distributions of plants with different adaptations for energy capture, together with large variation within sites in ground-layer light regime and among sites in soil texture, suggest that partitioning of light and soil gradients is important for maintaining the high plant diversity of oak savannas. Mass effects (involving dispersal subsidies from favorable microsites) may also play a role in maintaining diversity within these mosaic communities.