John F. Lehmkuhl

Modeling the effects of environmental disturbance on wildlife communities: avian responses to prescribed fire

Prescribed fire is a management tool used to reduce fuel loads on public lands in forested areas in the western United States. Identifying the impacts of prescribed fire on bird communities in ponderosa pine (Pinus ponderosa) forests is necessary for providing land management agencies with information regarding the effects of fuel reduction on sensitive, threatened, and migratory bird species. Recent developments in occupancy modeling have established a framework for quantifying the impacts of management practices on wildlife community dynamics. We describe a Bayesian hierarchical model of multi-species occupancy accounting for detection probability, and we demonstrate the models usefulness for identifying effects of habitat disturbances on wildlife communities. Advantages to using the model include the ability to estimate the effects of environmental impacts on rare or elusive species, the intuitive nature of the modeling, the incorporation of detection probability, the estimation of parameter uncertainty, the flexibility of the model to suit a variety of experimental designs, and the composite estimate of the response that applies to the collection of observed species as opposed to merely a small subset of common species. Our modeling of the impacts of prescribed fire on avian communities in a ponderosa pine forest in Washington indicate that prescribed fire treatments result in increased occupancy rates for several bark-insectivore, cavity-nesting species including a management species of interest, Black-backed Woodpeckers (Picoides arcticus). Three aerial insectivore species, and the ground insectivore, American Robin (Turdus migratorius), also responded positively to prescribed fire, whereas three foliage insectivores and two seed specialists, Clarks Nutcracker (Nucifraga columbiana) and the Pine Siskin (Carduelis pinus), declined following treatments. Land management agencies interested in determining the effects of habitat manipulations on wildlife communities can use these methods to provide guidance for future management activities.

Demography Of Northern Flying Squirrels Informs Ecosystem Management Of Western Interior Forests

We studied northern flying squirrel (Glaucomys sabrinus) demography in the eastern Washington Cascade Range to test hypotheses about regional and local abundance patterns and to inform managers of the possible effects of fire and fuels management on flying squirrels. We quantified habitat characteristics and squirrel density, population trends, and demography in three typical forest cover types over a four-year period. We had 2034 captures of flying squirrels over 41 000 trap nights from 1997 through 2000 and marked 879 squirrels for mark-recapture population analysis. Ponderosa pine (Pinus ponderosa) forest appeared to be poorer habitat for flying squirrels than young or mature mixed-conifer forest. About 35% fewer individuals were captured in open pine forest than in dry mixed-conifer Douglas-fir (Pseudotsuga menziesii) and grand fir (Abies grandis) forests. Home ranges were 85% larger in pine forest (4.6 ha) than in mixed-conifer forests (2.5 ha). Similarly, population density (Huggins estimator) in ponderosa pine forest was half (1.1 squirrels/ha) that of mixed-conifer forest (2.2 squirrels/ha). Tree canopy cover was the single best correlate of squirrel density (r = 0.77), with an apparent threshold of 55% canopy cover separating stands with low- from high-density populations. Pradel estimates of annual recruitment were lower in open pine (0.28) than in young (0.35) and mature (0.37) forest. High recruitment was most strongly associated with high understory plant species richness and truffle biomass. Annual survival rates ranged from 45% to 59% and did not vary among cover types. Survival was most strongly associated with understory species richness and forage lichen biomass. Maximum snow depth had a strong negative effect on survival. Rate of per capita increase showed a density-dependent response. Thinning and prescribed burning in ponderosa pine and dry mixed conifer forests to restore stable fire regimes and forest structure might reduce flying squirrel densities at stand levels by reducing forest canopy, woody debris, and the diversity or biomass of understory plants, truffles, and lichens. Those impacts might be ameliorated by patchy harvesting and the retention of large trees, woody debris, and mistletoe brooms. Negative stand-level impacts would be traded for increased resistance and resilience of dry-forest landscapes to now-common, large-scale stand replacement fires.

Ecological effects of alternative fuel-reduction treatments: Highlights of the National Fire and Fire Surrogate study (FFS)

The 12-site National Fire and Fire Surrogate study (FFS) was a multivariate experiment that evaluated ecological consequences of alternative fuel-reduction treatments in seasonally dry forests of the US. Each site was a replicated experiment with a common design that compared an un-manipulated control, prescribed fire, mechanical and mechanical + fire treatments. Variables within the vegetation, fuelbed, forest floor and soil, bark beetles, tree diseases and wildlife were measured in 10-ha stands, and ecological response was compared among treatments at the site level, and across sites, to better understand the influence of differential site conditions. For most sites, treated stands were predicted to be more resilient to wildfire if it occurred shortly after treatment, but for most ecological variables, short-term response to treatments was subtle and transient. Strong site-specificity was observed in the response of most ecosystem variables, suggesting that practitioners employ adaptive management at the local scale. Because ecosystem components were tightly linked, adaptive management would need to include monitoring of a carefully chosen set of key variables. Mechanical treatments did not serve as surrogates for fire for most variables, suggesting that fire be maintained whenever possible. Restoration to pre-settlement conditions will require repeated treatments over time, with eastern forests requiring more frequent applications.

Short-Term Responses of Red Squirrels to Prescribed Burning in the Interior Pacific Northwest, USA

Abstract We quantified changes in density of red squirrels (Tamiasciurus hudsonicus) in response to prescribed fire in mixed coniferous forests of Idaho and Washington, USA, using a Before-After-Control-Impact design. We found no evidence that low-severity prescribed fires affected density of red squirrels; we estimated the change in red squirrel densities due to prescribed fire as −0.15 squirrels/ha (95% CI u200a=u200a −0.405–0.105). Squirrel density did, however, increase with increasing live tree density, shrub cover, and density of large downed logs, and varied across years and states. These results indicate that land managers implementing prescribed fire treatments to reduce fuel loads on public lands can reduce the impacts of fire on squirrel populations by formulating prescriptions to retain large live trees and large downed logs.

Evaluating the effects of ecosystem management alternatives on elk, mule deer, and white-tailed deer in the interior Columbia River basin, USA

Abstract Elk (Cervus elaphus), mule deer (Odocoileus hemionus), and white-tailed deer (Odocoileus virginianus) are highly valued for their game, aesthetic, and spiritual qualities by sportsman, wildlife enthusiasts, and Native Americans in North America. As part of the Interior Columbia Basin Ecosystem Management Project (ICBEMP) of the US Forest Service and Bureau of Land Management, we: (1) defined key habitat associations of those species that could be used for a broad-scale (58 million hectares) analysis of management practices and (2) determined how three ecosystem management alternatives of a supplemental draft environmental impact statement (SDEIS) might affect the regional distribution of habitat for those species across the Basin over the next 100 years. For the three species, we developed a Bayesian Belief Network model that used available SDEIS datasets to estimate historical, current, and future habitat capability under the management alternatives in each of the 7467 subwatersheds (mean 8000xa0ha) in the study area. The model quantified “inherent habitat capability” as a function primarily of forage habitat capability, with cover as a minor influence. Forage habitat capability was a function of the percentage area of rangeland and early seral forest community types, and the qualitative influences of livestock overgrazing, wildfire, and prescribed fire. For the current and future periods, an “adjusted habitat capability” was estimated by adjusting inherent habitat capability for the negative effects of poor security from human disturbance. Open road density, cover area, and a terrain relief index were used to estimate the security effect. Habitat capability was reported by 15 ecological regions within the study area as mean subwatershed capability. Under all management alternatives, habitat capability increased about 5% for all three ungulate species over the next 100 years. Limitations of the coarse analysis scale restrict application of the model to large-scale assessments. Lacking regional population data, verification of model output was not feasible at the scale of analysis. However, the model was considered useful for tracking regional changes given the available habitat data and regional-scale objectives of the effort.

BUSHY-TAILED WOODRAT ABUNDANCE IN DRY FORESTS OF EASTERN WASHINGTON

Abstract We studied bushy-tailed woodrats (Neotoma cinerea occidentalis) in the eastern Washington Cascade Range to estimate their density and survival in 3 typical dry forest cover types. We predicted woodrat density to be high, moderate, and low in mature mixed-conifer forests, young mixed-conifer forests, and open ponderosa pine forests, respectively. We livetrapped on 8 × 8 grids (280 m) over an 8-day period each autumn for 4 years to obtain Huggins mark–recapture estimates of woodrat density. We captured woodrats 617 times and marked 193 individual woodrats in 12 sample stands during 42,165 trap nights. The sex ratio of captures was 62% female and 38% male. Adults made up 79% of captures. Woodrat density averaged 0.49 animals/ha among all cover types and years; woodrat density did not differ among types. Regardless of cover type, stands fell into 4 woodrat-density groups, with densities ranging from 0.13 to 0.93 woodrats/ha. Classification tree analysis revealed that woodrat-density groups could be predicted well (proportional reduction in error = 0.73 to 0.89) by the type and amount of cover provided by large snags, mistletoe brooms, and soft downed logs. Over the 4-year period, woodrat density ranged from 0.28 animals/ha to 0.87 animals/ha, increasing each year at a rate of λ = 1.8 consistently among cover types. A moderate (r = 0.55) density-dependent response in per capita rate of increase was detected. The apparent annual survival rate was a low 0.14. Management to reduce woody fuels and to restore low-intensity high-frequency fire regimes in ponderosa pine and dry Douglas-fir forest likely will reduce bushy-tailed woodrat populations unless prescriptions can mitigate the loss of snag, mistletoe, and downed log cover.

Breeding birds in riparian and upland dry forests of the cascade range

Abstract We quantified breeding bird abundance, diversity, and indicator species in riparian and upland dry forests along 6 third- to fourth-order streams on the east slope of the Cascade Range, Washington, USA. Upland dry forest on southerly aspects was dominated by open ponderosa pine (Pinus ponderosa) and dry Douglas-fir (Pseudotsuga menziesii) plant associations. Upland mesic forest on northerly aspects was dominated by closed-canopy Douglas-fir or dry grand fir (Abies grandis) plant associations. Riparian overstory vegetation was dominated by black cottonwood (Populus trichocarpa) plant associations with a prominent hardwood tree and shrub component. We quantified bird assemblages, diversity, and abundance from parallel point transects on riparian and adjacent dry and mesic upslope forests. We detected 80 bird species from >12,000 point-transect observations during 1998–1999. Eighteen species accounted for 75% of all detections. Species richness and evenness were similar in all 3 forest types, with approximately 35 species and high evenness (0.85) in each forest type. Bird species assemblages differed among dry, mesic, and riparian forest types, with the greatest differences between riparian and both dry and mesic upland forests. Riparian forest had the greatest number (9) of strong characteristic, or indictor, species among the 3 forest types. Upland mesic forest was characterized by 7 indicator species. Upland dry forest had 4 indicator species. Our results indicate that current standards and guidelines for riparian buffers zones would allow for avian refuge and corridor functions along these streams. Forest managers could use our indicator species to predict and monitor shifts in upland forest species composition from thinning and prescribed burning practices that are used to reduce fuels in uplands and to reduce continuity of fire effects between riparian and upland zones.

Mammal indicator species for protected areas and managed forests in a landscape conservation area of northern India

There is a realization that managed forests and other natural areas in the landscape matrix can and must make significant contributions to biodiversity conservation. Often, however, there are no consistent baseline vegetation or wildlife data for assessing the status of biodiversity elements across protected and managed areas for conservation planning, nor is there a rapid and efficient means to acquire those data. We used a unified vegetation classification and simple animal sampling design to describe the patterns of abundance of selected mammals as indicator, or characteristic, species in different vegetation types and protected areas vs. managed forest units in the Terai Conservation Area (TCA) in northern Uttar Pradesh state, India. We quantified the relative abundance of 15 mammals of conservation concern from dung counts in vegetation sampling plots within 122 sample patches in 13 vegetation types and 4 management units. Assemblages of species differed both among vegetation types and among management units. Species assemblages in the two protected areas differed strongly from those in two managed forests. Grasslands in protected areas were the most species diverse among vegetation types and had several indicator species. Protected forests were dominated by chital (Axis axis) and nilgai (Boselaphustragocamelus) in a second species group. A third species group in open grasslands and savannas in managed forests was characterized by cattle (Bostaurus) and Indian hare (Lepusnigricollis). Protected areas clearly are the core conservation area of the TCA for their relatively high habitat value and species diversity, and their protected status minimizes human disturbance. Impacts of human use are high in managed forests, indicating their compromised value for biodiversity conservation. Our simple assessment methodology gives managers a simple way to assess the status of important mammals across landscape conservation units.