Title: Small mammal and truffle response to burning in mixed-conifer forest
Investigators:
Marc Meyer, Malcolm North, USDA Forest Service, Sierra Nevada Research
Center, 2121 Second Street, Suite A-101, Davis, CA 95616; Douglas Kelt, Department of
Wildlife, Fish, and Conservation Biology, University of California, Davis 95616
Abstract
: There is little information to the intermediate (5 years or longer) effects of fuel reduction treatments on wildlife and their food webs. In previous research we found the abundance and diversity of truffles, a key food source in forest food webs, decreased as the intensity of fuel treatment increased. This short-term ( ≤ 2 years) response, however, may change as truffles, small mammal populations, and forest habitat rebound from prescribed fire and thinning treatments. Using a field experiment with fuel treatments completed in 2001, we will examine whether the intensity of fuel reduction influences this food web by measuring small mammal and truffle abundance, truffle consumption, and habitat variables in replicate mixedconifer plots 6-8 years following burning and thinning treatments. We will compare the response of the northern flying squirrel (a truffle specialist) and the lodgepole chipmunk (a dietary generalist) to assess how these ecologically distinct species respond to six different levels of fuel treatment. Changes in the abundance of these two species may significantly influence predators that rely on them, such as the California spotted owl, northern goshawk, and Pacific fisher. Forest managers working with the Teakettle Experiment have urged this comparison of longer-term changes in forest food chains as they work to implement fuel reduction treatments and understand their potential impact on sensitive species in California’s Sierra Nevada.
I. Introduction
Project Justification
Fuels are managed to reduce wildfire risk and enhance forest ecosystem health, yet it is not clear how different methods impact the forest food chain, animal abundance, or wildlife habitat. In the
Sierra Nevada, forest treatments are designed to reduce the risk of crown fires while simultaneously enhancing habitat conditions for the California spotted owl (
Strix occidentalis occidentalis
), yet the effect of different fuel reduction treatments on owl habitat quality is largely unknown. The spotted owl’s diet consists primarily of small mammals, particularly the northern flying squirrel (
Glaucomys sabrinus
), a species that relies on below-ground fungal fruiting bodies (‘truffles’) for most of its diet. Truffles also make up a substantial part of the diet of lodgepole chipmunks ( Neotamias speciosus ), an opportunistic consumer of fungi and prey for
California spotted owls, northern goshawks (
Accipiter gentilis
), and Pacific fisher (
Martes pennanti
) in the Sierra Nevada. Results from our previously funded JFSP research indicate that the abundance of truffles, frequency of truffle consumption, and number of flying squirrel captures decrease with increasing intensity of the fuel treatment. However, it is not clear how long these short-term ( ≤ 2 yrs.) results will persist over time.
We will examine how prescribed fire and mechanical thinning affect a food web that supports the California spotted owl and other species of management concern.
Project Objectives
Our research objective is to test whether the effect of fuel treatments on forest food webs is dependent on treatment intensity, such that high-intensity treatments (overstory thin, combined thin and burn) generally have a negative impact and low-intensity treatments (burn only, understory thin only) have a positive or neutral effect. Specifically, we predict the following mid-term responses:

1) truffle abundance and northern flying squirrel capture rates, habitat use, and truffle consumption will decline with increasing disturbance intensity (consistent with short-term patterns);
2) capture rates of lodgepole chipmunks, availability of non-truffle foods (e.g., seeds, fruits), and consumption of non-truffle foods by chipmunks will increase with disturbance intensity
(contrary to short-term patterns);
3) the burn-only and thin-only treatments will be similar to the control with respect to all food web variables (i.e., truffle abundance and consumption and small mammal capture rates; contrary to short-term patterns);
In response to requests from Sierra National Forest management personnel, this study will provide information on how different fuel reduction treatments affect the food web that supports the California spotted owl.
Background
Fire is an integral component of many ecosystems throughout the world (Dickman and
Rollinger 1998), and in forests it facilitates tree regeneration, nutrient cycling, and forest succession (Attiwill 1994). However, decades of fire suppression in many North American forests have altered these processes, increasing drought stress (Ferrell 1996), pest outbreaks
(Mattson and Hack 1987), and wildfire intensity (Husari and McKelvey 1996, Dickman and
Rollinger 1998). To reduce fuels and crown fire risk, two treatments often are employed: mechanical thinning alone in wildland urban interface areas, and thinning and prescribed fire in more remote locations. It is unclear, however, how thinning, fire, and their interaction influence the ecological processes and trophic structure that maintain ecosystem ‘health’ and function.
An important functional component of forest ecosystems are ectomycorrhizal fungi
(EMF), which are required by most temperate forest conifer species for increased water and nutrient uptake (Molina et al. 1992). EMF fruiting bodies, especially belowground truffles, form the base of a complex food web in forests (Johnson 1996) and are a major part of the diet of many mammals (Fogel and Trappe 1978, Maser et al. 1978). Truffles are the primary food of northern flying squirrels (
Glaucomys sabrinus
; Pyare and Longland 2001a), which in turn constitute the main prey of California spotted owls ( Strix occidentalis occidentalis ) in Sierra
Nevada mixed-conifer forests (Williams et al. 1992). In contrast, many species of chipmunks
(
Neotamias
spp.) are opportunistic consumers of fungi, consuming a high frequency of truffles throughout the summer and fall when availability is high (Pyare and Longland 2001a, Meyer et al. 2005a). The lodgepole chipmunk (
Neotamias speciosus
), a species sympatric with northern flying squirrels throughout much of the Sierra Nevada, opportunistically consumes fungi, seeds, vegetation, arthropods, and lichens (Tevis 1953, Best et al. 1994, Meyer et al. 2005a). Both small mammals and truffles comprise a significant part of the diet of the fisher ( Martes pennanti ) and American marten (
Martes americana
; Zielinski and Duncan 2004), key management species in Sierra Nevada forests. Viable fungal spores pass through the digestive system of mammal consumers and are dispersed to new soil patches where they facilitate conifer succession
(Terwilliger and Pastor 1999) and promote EMF diversity (Johnson 1996, Izzo et al. 2005).
Since thinning and burning have fundamentally different impacts on forest overstory and understory structure, these two fuel reduction methods should have different effects on truffleproducing EMF and truffle-consuming mammals. Recent studies predict that the abundance of
EMF and richness of truffle species peaks in areas with well-developed surface litter and organic material (Amaranthus et al. 1994, Claridge et al. 1999, Lehmkuhl et al. 2004) and a higher density of large-diameter trees with greater canopy closure (States and Gaud 1997, Lehmkuhl et

al. 2004). Burning can directly impact truffle-producing EMF through heat stress (Dahlberg et al. 2001, Smith et al. 2005) and indirectly by reducing surface litter (Waltz et al., 2003), log cover (Knapp et al., 2005), and herbaceous plant or shrub cover (Meyer et al., 2005b).
Reduction in canopy cover and increased bare ground conditions immediately following burning can increase surface soil temperatures and lead to suppressed truffle productivity (Meyer et al.
2005b). Mechanical thinning may impact overstory features such as canopy cover and large trees (Waters and Zabel, 1995; Waltz et al., 2003), but can have less impact on understory features associated with truffles ( e.g.
, litter and decayed logs; Meyer et al., 2005b). Both burning and thinning can reduce canopy, shrub, and log densities that provide protective cover and nest sites for truffle-consuming small mammals (Meyer et al. 2005b).
Results from our previously completed JFSP project indicate that the abundance of truffles, frequency of truffle consumption, and number of flying squirrel captures all decrease with increasing intensity of the fuel treatment (Meyer et al. 2005b; Meyer et al., in review
). In particular, when management treatments are used in combination (i.e., mechanical thinning followed by burning) the impact on forest food webs is greater than either treatment alone (i.e. either burning or thinning; Meyer et al. 2005b). Additionally, intensive overstory thin treatments have a greater impact on truffle abundance and consumption than low-intensity understory thin treatments (Meyer et al. 2005b). These short-term results suggest food webs are immediately impacted by treatment intensity, but longer-term ecosystem recovery is still unknown.
We predict that food web responses will change after a rapid initial response to treatments, as forest understory and overstory conditions stabilize. Immediately after a fire, logs are reduced, litter is removed, and host trees are stressed; these conditions reduce EMF abundance and may impact truffle production and consumption in the short term (Smith et al.
2005; Meyer et al. 2005b). These immediate responses, however, will be offset as forest regeneration processes lead to enhanced EMF and small mammal colonization. As trees die, litter and log inputs will gradually increase organic material associated with truffle production
(Amaranthus et al. 1994) and provide refugia and foraging sites for small mammals (Pyare and
Longland 2001b, 2002; Meyer et al., in press ). Increases in canopy and understory cover can increase soil moisture, organic material, and buffer soil microclimates, thereby enhancing root growth, EMF biomass, and truffle production (North et al. 1997, Smith et al. 2005, Meyer and
North 2005). Increased truffle abundance and canopy cover can, in turn, attract small mammal consumers and enhance truffle consumption. Truffle-specialists, such as the northern flying squirrel, may become increasingly attracted to post-burned or -thinned microhabitats where denser canopies and thicker organic layers have developed (Meyer et al., in review ). In contrast, dietary-generalists, such as the lodgepole chipmunk, may become increasingly attracted to opencanopy post-treatment stands where abundant understory vegetation (i.e., shrubs, herbs) provide ample food and protective cover. Although it is not clear how rapidly these ecosystems will rebound from management treatments, our initial study suggests changes in stand structure have rapid, cascading effects on forest food webs (Meyer et al. 2005b).
The purpose of this study is to examine the mid-term effects (5-7 years) of fire and thinning on the abundance and microhabitat associations of truffles and small mammals as well as the consumption of truffles by small mammals in mixed-conifer forest of the southern Sierra
Nevada of California. Specifically, we will evaluate the mid-term effects of burning, two levels of thinning, and thinning followed by burning, on: (1) the frequency, biomass and species richness of truffles, (2) small mammal capture rates and microhabitat use, and (3) frequency and generic richness of truffles consumed by two groups of small mammals, a truffle-specialist
(northern flying squirrel) and dietary generalist (lodgepole chipmunk). We also will examine changes in forest stand structures (e.g., canopy cover, litter depth, tree density) that are

associated with truffle abundance following management treatments (e.g., North and Greenberg
1998, Lehmkuhl et al. 2004). The range of forest conditions produced by the six treatments in our study allow for strong inference about which forest habitat conditions are associated with rebounding food webs.
II. Materials and Methods
Study area
The study will be conducted at Teakettle Experimental Forest, a 1300-ha, mixed-conifer forest in the southern Sierra Nevada, Fresno Co., California. Teakettle Experimental Forest
(1800-2400 m elev.), characterized by a Mediterranean-influenced montane climate, with hot, dry summers, receives precipitation almost exclusively as snow in the winter (Major 1990).
Mean annual precipitation is 125 cm at 2100 m, and mean summer rainfall (June-September) during 1999-2002 was 0.9 ± 0.3 cm. Teakettle Experimental Forest is an old-growth forest characterized by a multi-layered canopy and numerous large (>100 cm diameter at breast height; dbh) and old (>200 yrs.) trees, snags, and decayed logs. Dominant trees are white fir (
Abies concolor
(Gord & Glend.) Lindl.), red fir (
A. magnifica
A. Murray), sugar pine (
Pinus lambertiana Douglas), Jeffrey pine ( P. jeffreyi Balfour), and incense cedar ( Calocedrus decurrens
(Torrey) Florin).
Experimental design and treatments
In 1997, the USDA Forest Service Pacific Southwest Research Station and the Sierra
National Forest initiated an Administrative Study called the Teakettle Ecosystem Experiment.
Figure 1: Location of the Teakettle Experimental Forest, layout of the 18 plots (4 ha each, 3 replicates of each treatment) and sampling grid used within each plot. Plot codes indicate

treatment (described below) and replicate number.
200 m
The Teakettle Experiment used a full factorial set of treatments in a completely randomized design. Two levels of burning (U-unburned/B-prescribed burned) were crossed with three levels of thinning (N-none, C-understory thin and S-overstory thin) producing six treatments (Figure 1). Within Teakettle Experimental Forest, 18 replicate 4-ha plots were established and all plots but one were randomly assigned to each treatment. The single exception was a plot that was randomly assigned to one of the three unthinned treatments because it contained a perennial stream around which Forest Service regulations did not permit tree harvest.
All plots are separated by untreated buffer zones of 50-150 m. The size and spatial placement of plots were determined following variogram and cluster analysis of mapped sample quadrats
(North et al. 2002). In July-September of 2000 and July-August of 2001, twelve plots were experimentally thinned following two prescriptions: understory thinning following CASPO
(California Spotted Owl) guidelines and overstory thinning following a shelterwood prescription.
Under CASPO thinning, no trees ≥ 76 cm diameter were harvested and at least 40% of the canopy cover was left in place after harvest. Although initially designed for minimizing impact to spotted owl habitat, the CASPO guidelines became the standard forest practice in the 1990s and are still widely used as a fuel reduction treatment (SNFPA 2004). With shelterwood thinning, all trees >25 cm dbh were removed except for approximately 22 dominant, evenlyspaced trees per ha. Shelterwood guidelines followed practices used on several National Forests

in the Sierra Nevada before CASPO regulations and are considered a more aggressive fuel treatment that leaves widely spaced tree canopies more resistant to crown fire. In early
November 2001, after the first substantial fall rain, twelve plots were prescribed burned. At this time, average daytime temperature was 13 ° C and relative humidity ranged 25-70%. The percentage of ground cover burned was approximately 35-70% and 20-40% in thinned and unthinned plots, respectively.
Truffle sampling
In June and August of 2000 and June of 2001, pre-treatment frequency, species richness, and biomass of truffles were estimated for each treatment plot by placing five 4-m 2 circular quadrats immediately outside the treatment plots (to minimize pre-treatment soil disturbance).
Nine sample points for long-term monitoring were established within each plot in a 3 by 3 grid with 50 m spacing between points and a 50 m buffer from the plot boundary. Beginning in
August 2006, a single 4 m 2 circular quadrat will be placed at the first eight grid points per plot
(including one quadrat at the ninth grid point in one of the three plots), giving a total sample area of 32 m 2 per plot (36 m 2
32 + 32 + 36 m 2
for the third replicate), 100 m
), and 600 m 2
2 per treatment (3 replicates per treatment;
total per season (6 treatments). In August 2006 and June and
August of 2007 and 2008, quadrats will be sampled for truffles by searching through the litter, humus, and upper 5 cm of mineral soil using a four-tined rake. All collected truffles will be counted, placed in wax bags, dried for 24 hours at 60 ° C, weighed to the nearest 0.01 g, and identified to species. Truffle collections will be used to estimate frequency, species richness, and total biomass of truffles in treatment and control plots.
Following Waters et al. (1997), secotioid, epigeous fungi (
Hymenogaster
and
Martellia
) will be grouped with truffles in fungal collections, because these taxa are mycorrhizal and primarily producers of subterranean fruiting bodies. We will use keys by Smith (1966), Smith and Smith (1973), Smith and Zellner (1966), and Arora (1986) to identify species. The peridium, gleba, columella, and spores of fresh specimens will be examined microscopically, tissues reinflated with 3% potassium hydroxide, and Melzer’s reagent (I, K, and chloral hydrate;
Castellano et al. 1989) will be used to characterize dextrinoid (reddish brown) and amyloid
(blue-black) reactions. Samples will also be compared to an extensive collection of voucher specimens (878 individuals of 87 species) collected from previous studies at Teakettle (Meyer and North 2005, Meyer et al. 2005b) and a nearby 1 ha sampling site (North 2002).
Small mammal sampling and truffle consumption
Lodgepole chipmunks and northern flying squirrels will be censused in all treatment plots with Tomahawk live traps (model 201; 40 x 13 x 13 cm). Traps will be attached 1.5 m high on the trunk of a large (>70 cm dbh) tree at each 50 m grid point in a plot ( n = 9 traps per plot,
Figure 1). In 2007-2008, traps will be placed in each plot for three consecutive days in early summer (June) and four consecutive days in mid-summer (July-August), for a total of 63 trapnights per plot per year. Traps will be baited with a mixture of peanut butter and rolled oats, and a small cardboard shelter filled with polyester stuffing material will placed in each trap to provide animals with thermal insulation. Captured animals will be uniquely marked with numbered ear tags, and standard measurements (weight, gender, reproductive condition) will be recorded. All captured animals will be handled in accordance with the American Society of
Mammalogists Animal Care and Use Committee (1998) guidelines and an approved institutional animal and care use committee (IACUC) protocol.

To determine what small mammals are eating, fresh fecal samples will be opportunistically collected from captured lodgepole chipmunks and northern flying squirrels.
Samples will be placed into individual envelopes, labeled, and stored in a dry location at room temperature for 1-6 weeks. Following the methods of Colgan et al. (1997) and Pyare and
Longland (2001a), small portions of all pellets in a single fecal sample (approximate total weight
= 25 mg) will be assessed for truffle spores and generalized plant material. Samples will be mixed with 3 ml of distilled water and a single drop is applied to each of three slides. One drop of Melzer’s reagent will be added to each slide, and 25 systematically located fields of view will be examined at 400× magnification (total of 75 fields of view per sample). All fungal spores present will be identified to genus using a synoptic key (Castellano et al. 1989). Frequency of occurrence will be calculated separately for each genus. The frequency of occurrence of nonsecotioid, epigeous genera, generalized plant material (e.g., leaves, fruits, seeds), and arthropods also will be recorded. Fecal samples collected from individuals that are captured in more than one treatment plot will be excluded from analysis.
Measurement of stand variables and cone abundance
At each grid point per plot, we will collect data on nine stand variables within a 12.6 m radius (0.05 ha) of each station: (1) total basal area of large (>50 cm dbh) trees, (2) total number of large (>50 cm dbh) snags, (3) canopy cover (%), (4) total volume (m 3 /ha) of intact (decay classes 1-2; Cline et al
., 1980) and heavily decayed logs (decay classes 3-5), (5) litter depth (cm),
(6) distance to shrub cover (m), (7) herbaceous plant cover (%), (8) aspect, and (9) distance to perennial creek (m). We chose these variables based on published studies of northern flying squirrels (Carey, 1995; Waters and Zabel, 1995; Carey et al
., 1999; Cotton and Parker, 2000;
Pyare and Longland, 2001a, 2002; Bakker and Hastings, 2002; Ford et al
., 2004; Smith et al
.,
2004), lodgepole chipmunks (Best et al. 1994), and truffles (Fogel 1976, Luoma et al. 1991,
Amaranthus et al. 1994, Clarkson and Mills 1994, States and Gaud 1997, Waters et al. 1997,
North and Greenberg 1998, Lehmkuhl et al. 2004, Meyer and North 2005). In each 12.6 m plot, the dbh of all trees and snags >50 cm dbh will be measured. Canopy cover will be estimated using hemispherical photographs that will be analyzed using Gap Light Analyzer 2.0 (Simon
Fraser University, Burnaby, British Columbia, Canada) software. Litter depth will be measured by digging 3 shallow pits at the edge of each quadrat (at 0, 120, and 240 ˚ ) and taking three depth measurements of the combined litter and humus layers. Herbaceous plant cover will be visually estimated using standard sampling procedures established at Teakettle in 1998 (North et al. 2005;
Wayman and North in review). Aspect will be determined using a compass, and distance to perennial creek will be measured with a laser rangefinder.
Conifer cone abundance will be sampled in all 4-ha plots to determine potential relationships between food availability and chipmunk variables (capture rate, reproductive condition) across treatments and years. Relative conifer cone abundance will be assessed using trees that have been annually surveyed at Teakettle since 1998. We will conduct two surveys
(July and October) of the four most common overstory tree species (
A. concolor
,
A. magnifica
,
P. jeffreyi , and P. lambertiana ) that provide nutritious seeds. The July and October survey provides an estimate of summer and over-winter cone availability, respectively. An overstory tree (>80 cm dbh) of each of the four species within each of 18 treatment plots has been tagged
(n = 72). In July and October of 2007 and 2008, the total number of cones on each tree will be counted using 10 × 40 binoculars at a monumented survey station located at least 35 m from the tree base. Surveys tally only observable cones and therefore cannot estimate total cone abundance. However by repeatedly sampling from the same observation station annual differences between observed stand crops can be compared (Meyer et al., in review ). We also

infer changes in cone abundance due to treatment by standardizing annual tallies using each tree’s pre-treatment mean abundance. To compliment the cone surveys, we will also use 9 systematically placed 0.25 m 2 seed traps in each of the 18 plots. These were established in 2001 and each year, seeds and litter are collected, sorted, identified to species, dried, and weighed
(Ruth Kern, Fresno State University). These two methods will give relative but consistent measures of arboreal cone and surface seed abundance.
Analysis
In our previous study all pre-treatment truffle variables (frequency, species richness, biomass, dietary generic richness, and dietary frequency) were analyzed with 3-way mixedmodel analysis of variance (ANOVA) to test for differences among treatments (fixed factors) and between pre-treatment years (2 years; random factor). There were no significant differences
(
P
> 0.05) in any variables among pre-treatment factors.
Post-treatment data will be evaluated for normality with the Kolmogorov-Smirnov test and for homoscedasticity with Levene’s test (Zar 1984). Multivariate analysis of variance
(MANOVA) and 2-way Model 1 ANOVAs will be used to test for the effect of fire (2 levels), thinning (3 levels), and post-treatment year (3 years) on truffle production (species richness, frequency, and biomass) and consumption by flying squirrels and lodgepole chipmunks (dietary generic richness and frequency). A single MANOVA will be used to test the null hypothesis that low-intensity (CASPO) and high-intensity (shelterwood) thinned plots were similar with respect to dietary generic richness and frequency of truffles. Additionally, a single ANOVA will be used to examine whether low- and high-intensity thinned plots were similar with respect to the frequency of non-truffle food items (e.g., vegetation) in the diets of lodgepole chipmunks.
MANOVAs will be used to evaluate whether dietary generic richness and frequency were similar between burned plots, and both thinned and thinned plus burned plots. Two-factor Model 1
ANOVAs will be used to examine if the three most common truffle genera in diet samples are similar among plots both burned and thinned and those exclusively burned. Tukey’s HSD test will be calculated for all stand variables and dietary truffle taxa, as well as truffle frequency, biomass, species richness, dietary frequency, and dietary generic richness to evaluate responses of specific variables to burning and thinning treatments. In addition, 95% confidence intervals will be calculated for truffle frequency, biomass, species richness, dietary frequency, and dietary generic richness.
Indicator species analysis (ISA; Dufrene and Legendre 1997) using species of truffles will be used to identify significant species associations with burning and thinning treatments.
ISA combines information on relative abundance and site fidelity of each species to estimate indicator values (
IV
) for each species in each group. ISA
P
-values will be calculated as the proportion of 1000 randomized trials with an
IV
equal to or greater than the observed
IV
:
P
= (1
+ number of runs ≥ observed)/ (1 + number of randomized runs).
We will select all stations where flying squirrels are captured in June and August and an equal number of random stations where northern flying squirrels are not captured (i.e., an equal number of capture and no-capture stations per treatment) for microhabitat analysis. We will use
Akaike’s Information Criterion (AIC) to reduce our original set of microhabitat variables to avoid model over-fitting (Burnham and Anderson, 1998). We will use a corrected AIC (AIC c
) for model selection, since sample sizes will be small relative to the number of model parameters
(Burnham and Anderson, 1998). The selected logistic regression will relate all selected microhabitat variables to the presence or absence of flying squirrels and lodgepole chipmunks.
For each significant parameter in the logistic regression model, we will calculate odds ratios and their confidence intervals based on a Quasi-Newton estimation that approximates the second-

order derivatives of the retrospective loss function (Statistica, 2003). The odds-ratio estimates will be interpreted as the odds of capturing a flying squirrel or chipmunk given a one unit change in a microhabitat parameter ( e.g.
, litter depth) after being adjusted for the effects of other microhabitat parameters in the model (Smith et al., 2004). We will use a sensitivity analysis of each treatment type to evaluate the performance of the reduced logistic regression model and assess model accuracy in successfully predicting captures of flying squirrels and chipmunks among trap stations (Hosmer and Lemeshow, 2000).
Since the number of captures and recaptures varied greatly among treatments during our earlier JFSP project, we will calculate capture rate as the relative abundance of flying squirrels and chipmunks among treatments (cf. Waters and Zabel 1998). Capture rate (CR) will be expressed as a modification of the formula of Nelson and Clark (1973): CR = [I/(T – S/2)] × 100, where I = number of individuals captured, T = number of traps multiplied by the number of nights traps were opened (e.g., trapnights), and S = number of traps sprung across all nights of effort. Capture rate is expressed as a percentage of traps with ≥ 1 capture and is positively correlated to Nelson and Clark’s (1973) capture rate values (Pearson’s R
2 = 1.000 and 0.972 for
2002 and 2003, respectively). We also will calculate the total number of individuals captured, as these counts were proportional to population size in small mammal populations (Slade and Blair
2000). Multivariate analysis of variance (MANOVA) and Model I 2-way ANOVAs will be used to test for the effect of fire (2 levels) and thinning (3 levels) on chipmunk capture rates and demographic variables. A single MANOVA will be used to examine the effect of these treatments on flying squirrel and chipmunk sex ratio (of all individuals), percent of females in breeding condition (i.e., females lactating or pregnant), and male and female weight.
The total number of conifer (
A. concolor
,
A. magnifica, P. jeffreyi
, and
P. lambertiana
) cones will be calculated as the total number of cones per large ( ≥ 50 cm dbh) tree × the number of large trees per hectare in each plot. The total weight of seeds by species and treatment will be determined from annual collections from the seed traps. Two-factor ANOVAs will be used to examine the effect of burn and thin treatments on the total number of cones for each of the four conifer species and on total seed biomass by treatment. Single-factor ANOVAs will be used to examine differences in the number of conifer cones per tree, seed biomass, and understory plant coverage (shrubs, herbaceous plants) between 2007 and 2008. Pearson’s product-moment correlation will be used to examine the relationships between the availability of food resources
(
A. concolor
,
A. magnifica
,
P. lambertiana
, and
P. jeffreyi
cone abundance, seed biomass, coverage of shrubs and herbaceous plants) and chipmunk capture rate or body mass of females.
Science Delivery and Application
The Teakettle Experiment was initiated in 1997 in response to management concerns over how fuels treatments affect ecosystem ‘health’. In truth, the experiment’s science delivery and application was slow and directionless at first as both scientists and managers struggled to bridge cultural differences between agencies. After 10 years, however, persistence has produced an efficient collaboration where management helps establish research priorities and scientists strive to provide relevant results within the context of good experimental design. This proposal was developed following requests from Sierra National Forest personnel for information on longterm response of sensitive species (California spotted owl, northern flying squirrel and Pacific fisher) to different fuel treatments. The goal of the Teakettle Experiment has always been to work directly with managers, and we expect to continue this relationship, communicating results to federal and state agency personnel.
Specifically this project’s outreach will work in three areas:

(1)
(2)
(3)
Agency workshops – Teakettle will institute annual workshops, which will include fire officers, silviculturists and wildlife biologists. We believe this may be the best means of getting interdisciplinary ecological information to forest managers and scientists at the same time so that cooperation can be initiated from within an agency.
Public outreach - Science and management information has and will continue to be presented to the general public, and education and management organizations (UC Davis, California
State University-Sacramento, CAFE, USDA, Environmental Science Academy of Merced
High School District). This has included presentations at UC Davis, King’s River
Administration Study Symposium (FS conference), Yosemite National Park, high schools, the local chapters of the Sierra Club and National Audubon, and field sight visits from interested groups, which have included the Quincy Library Group, California Timber
Association and USFS personnel from the Washington office. Teakettle has been designated as a demonstration site by the Joint Fire Science Program and we will regularly communicate with interested public groups.
Publications - Scientific information will be presented in peer reviewed publications and presentations of papers at scientific symposia and meetings. A total of four peer-reviewed publications (in addition to 3 manuscripts currently in review) and numerous scientific presentations have resulted from our previously completed JFSP project.
Expected Benefits of the Proposal
Managing fuels treatments is an immediate focus of many forestry practices, but the public is often very concerned about potential impacts on wildlife habitat and forest health. Very little information is available about how different fuel treatment prescriptions affect forest food webs and how these changes impact wildlife abundance and habitat use. Much of the controversy over fuels treatments has focused on the perceived negative effects of fire and thinning disturbances on wildlife, despite the fact that these forests had a historic disturbance regime of frequent, low-intensity fire. Immediate effects of fuel treatments are likely to reduce wildlife abundance as animals adjust to new forest conditions. This immediate response may reinforce perceptions that fuel treatments are disruptive of a delicate balance supporting wildlife and their habitat. A longer, intermediate-term response, however, may provide important information regarding the resiliency of forest food webs to different treatments that approximate or depart from historic disturbances to which wildlife are adapted.
Science and management information from this project has and will continue to be presented to the general public, and education and management organizations. This research will benefit USDA Forest Service fire and land management personnel through presentations and reports as well as researchers at the Pacific Southwest Research Station, University of California
Davis, and other academic and research institutions.
Literature Cited
Attiwill, P. M. 1994. The disturbance of forest ecosystems: the ecological basis for conservative management. Forest Ecology and Management 63:247-300.
Claridge, A. W., S. C. Barry, S. J. Cork, and J. M. Trappe. 1999. Diversity and habitat relationships of hypogeous fungi. II. Factors influencing the occurrence and number of taxa. Biodiversity and Conservation 9:175-198.
Dickman, D. I., and J. L. Rollinger. 1998. Fire for restoration of communities and ecosystems.
Bulletin of the Ecological Society of America 79:157-160.
Fogel, R., and J. M. Trappe. 1978. Fungus consumption (mycophagy) by small mammals.
Northwest Science 52:1-31.

Husari, S. J., and K. S. McKelvey. 1996. Fire management policies and programs. In Sierra
Nevada Ecosystem Project, final report to congress, vol. 2, assessment and scientific basis for management options. Wildlands Center Research Report, Davis, California. pp.
1101-1118.
Izzo, A. D., M. Meyer, J. M. Trappe, M. North, and T. D. Bruns. 2005. Hypogeous ectomycorrhizal fungal species on roots and in small mammal diet in a mixed-conifer forest. Forest Science 51:243-254.
Johnson, C. N. 1996. Interactions between mammals and ectomycorrhizal fungi. Trends in
Ecology and Evolution 11:503-507.
Lehmkuhl, J. F., L. E. Gould, E. Cazares, and D. R. Hosford. 2004. Truffle abundance and mycophagy by northern flying squirrels in eastern Washington forests. Forest Ecology and Management 200:49-65.
Meyer, M. D., D. A. Kelt, and M. North. 2005a. Fungi in the diets of northern flying squirrels and lodgepole chipmunks in the Sierra Nevada. Canadian Journal of Zoology 83:1581-
1589.
Meyer, M. D., D. A. Kelt, and M. North. Microhabitat associations of northern flying squirrels in burned and thinned forest stands of the Sierra Nevada. In press in
American Midland
Naturalist .
Meyer, M. D., M. North, and D. A. Kelt. 2005b
.
Short-term effects of fire and forest thinning on truffle abundance and consumption in the southern Sierra Nevada of California. Canadian
Journal of Forest Research 35:1061-1070.
Molina, R., H. B. Massicotte, and J. M. Trappe. 1992. Specificity phenomena in mycorrhizal symbioses: community-ecological consequences and practical implications.
In
Mycorrhizal functioning: an integrative plant-fungal process.
Edited by
M. Allen.
Chapman Hall, New York. pp. 357-423.
North, M., J. Trappe, and J. Franklin. 1997. Standing crop and animal consumption of fungal sporocarps in Pacific Northwest forests. Ecology 78:1543-1554.
Pyare, S., and W. S. Longland. 2001a. Patterns of ectomycorrhizal-fungi consumption by small mammals in remnant old-growth forests of the Sierra Nevada. Journal of Mammalogy
82:681-689.
Smith, J. E., D. McKay, G. Brenner, J. McIver, and J. W. Spatafora. 2005. Early impacts of forest restoration treatments on the ectomycorrhizal fungal community and fine root biomass in a mixed conifer forest. Journal of Applied Ecology 42:526-535.
Waters, J. R., K. S. McKelvey, C. Zabel, and W. W. Oliver. 1994. The effects of thinning and broadcast burning on sporocarp production of hypogeous fungi. Canadian Journal of
Forest Research 24:1516-1522.