1
Table of Contents
Contributors
3
Introduction
4
Purpose of Study
5
Objectives of Study
5
Research Highlights
6
Summary
7
Response Variable Modules
Chapter 1: Fuels and Fire Module
9
Chapter 2: Terrestrial Bird Module
33
Chapter 3: Spotted Owl Module
75
2
Contributors:
Fire and Fuels Module:
Dr. Brandon Collins, Research Forester, USDA Forest
Service, Pacific Southwest Research Station, Davis, CA
Phone: 530-759-1701 Email: bmcollins@fs.fed.us
Christopher Dow, Biological Science Technician, USDA Forest
Service, Pacific Southwest Research Station, Davis, CA
Gary Roller, Biological Science Technician, USDA Forest
Service, Pacific Southwest Research Station, Davis, CA
Dr. Scott Stephens, Associate Professor of Fire Science
Ecosystem Sciences Division, University of California, Berkeley
Terrestrial Bird Module:
Ryan D. Burnett, PRBO Conservation Science
Phone: 530-258-2869
Email: rburnett@prbo.org
Paul Taillie, PRBO Conservation Science
Nathaniel Seavy, PRBO Conservation Science
Spotted Owl Module:
PLAS / PSW
Research Coordination:
Dr. John Keane, Research Ecologist, USDA Forest Service, Pacific
Southwest Research Station, Davis, CA
Phone: 530-759-1704 Email: jkeane@fs.fed.us
Dr. Mary Conner, Research Ecologist/Biostatistician, USDA
Forest Service, Pacific Southwest Research Station, Davis, CA
Claire V. Gallagher, Wildlife Biologist, USDA Forest Service,
Pacific Southwest Research Station, Davis, CA
Dr. Ross A. Gerrard, Research Geographer, USDA Forest Service,
Pacific Southwest Research Station, Davis, CA
Gretchen Jehle, Wildlife Biologist, USDA Forest Service
Pacific Southwest Research Station, SNRC
Paula A. Shaklee, Wildlife Biologist, USDA Forest Service
Pacific Southwest Research Station, SNRC
Dr. Rick Bottoms, Program Manager, USDA Forest Service, Pacific
Southwest Research Station, Davis, CA.
Phone: 530-759-1703 Email: rmbottoms@fs.fed.us
3
Introduction:
The Pacific Southwest Region and the Pacific Southwest Research Station agreed in 2002 to
jointly develop and fund an administrative study to fill management information needs
concerning the relationship between management-caused changes in vegetation and their
effects on spotted owl habitat and population dynamics. The detailed discussions explaining
how this program was started are provided in previous Annual Reports. Copies of previous
Annual Reports for this program are available on the Sierra Nevada Research web site
(http://www.fs.fed.us/psw/topics/ecosystem_processes/sierra/forest_health/pls/) or upon
request.
This is the tenth such Annual Report that we have compiled. The primary purpose of this is
to provide a periodic synopsis of what we have been learning so all interested parties can
remain informed of the progress. Research products resulting from this effort will be
disseminated as they are ready and this will vary from module to module, project to project,
and from year to year. We expect that there will be a continuous flow of findings
documented primarily with publications in both refereed journals and other publication
outlets. The cadre of scientists, support staff, students, and others contributing to this effort
will also be making oral presentations and providing other kinds of outreach materials to
help inform interested parties and our peers on the results of this work.
We provide some review information here to reinforce the intent of our work. This background
information provides a general overview on the purpose of this research program and helps set
the context for the report. We have emerged from the initiation phase and we have collected
an impressive amount of information. Many publications have been completed and some key
ones are in development. We expect to provide useful information, particularly from the
spotted owl module, in the immediate future. Of course much of our research purpose depends
on forest management treatments to be put in place and then observing short and even long
term response to those treatments. In several areas treatments have been completed.
Observations of response after treatments will logically take place in the ensuing years. We
anticipate that the main body of field research will conclude by the end of Fiscal Year 2012.
New studies are being contemplated and we may collectively, in collaboration with land
managers, choose to extend facets of this work beyond this date.
We recognize that response of different elements of the forest can occur immediately after
treatments; however, it is also possible that response can occur slowly and not be recognized
for some period of time depending on the response variable of interest. Alternatively it is also
possible that some response variables exhibit a notable initial response and then return to a
state similar to that of before the treatments. Thus we believe it is prudent to look at a fairly
long period of post treatment response if possible, even if funding limitations require
scheduling follow-up work in stages over time with periods of inactivity.
4
Purpose of the Study:
This Plumas Lassen Study (PLS) is interdisciplinary by design, examining at least five groups
of response variables (spotted owls, small mammals, terrestrial birds, vegetation, and fuels
conditions) through collaboration between researchers of the USDA Forest Service Pacific
Southwest Research Station (PSW) and cooperators from the Universities of California,
Berkeley and Davis, and PRBO Conservation Science. The study addresses some of the most
significant uncertainties that confound management decisions in the Sierra Nevada today,
including in the Herger-Feinstein Quincy Library Group (HFQLG) Pilot Project Area. How do
old-forest-dependent species respond to vegetation management over space and time? Do fuels
management approaches effectively address fuels loadings without negatively affecting species
viability? How effective are landscape level fuels management strategies in modifying fire
behavior and reducing the extent and severity of wildland fire? These and related questions are
the focus of the work being done in this study.
Objectives of Study:
The original overarching objective of this research was to address an array of related
ecological questions in a coordinated, integrated effort, thereby providing empirical data to
inform future management decisions. The landscape scale of this design was both the
driving force addressing the key questions as well as the largest impediment to successful
construction of a scientifically credible experimental design and implementation in the field.
Our research team believes that assessing many of the key elements of forest ecosystems
should be done over larger spatial and temporal scales than has typically been investigated
in past research. The important difference we are investigating is the response to changes in
forest structure and composition over space and time rather than simply site specific and
immediate response. We believe this difference is especially relevant to forest management
practices that are designed for large landscapes, executed over relatively long time frames,
such as landscape level fuels treatment strategies.
This research program is designed to address the three principal issues described below.
These issues are specifically addressed through research questions and attending
investigational approaches tailored for five different research components of this research
program. These specific questions are detailed in the individual study plans for each module.
Here we simply highlight the main objectives of the integrated research program and
summarize the primary research questions that we plan to pursue.
Wildland Fire Behavior and Protection. How do landscape level fuels and
silvicultural treatments affect potential fire behavior and effects? Are specific
combinations of defensible fuel profile zones (DFPZs) and subsequent individual
tree selection or area treatments to thin the matrix effective in reducing the extent
and severity of wildland fires? Are realized fire management benefits consistent
with hypothesized results in reducing fire risk and altering fire behavior?
Landscape Dynamics. How do combinations of DFPZs, subsequent individual
tree selection or area treatments to thin the matrix, group selection, riparian
5
protection standards, and species-specific protection measures affect landscape
dynamics such as forest structure, composition, and succession at multiple scales
of space and time?
Species Viability. How will old-forest dependent species, particularly the
California spotted owl (CSO) (Strix occidentalis occidentalis) and its prey base,
comprised of various species of small mammals, respond to management-induced
changes in vegetation composition, structure, and distribution over space and
time? How is response to treatments manifested at the individual and population
levels of biological organization?
Research highlights for 2011:
Funding that supported the 2011 PLS research came from both the HFQLG Pilot Project and the
Storrie Fire Restoration effort. The combined funding has shifted research priorities to
specifically address restoration considerations with respect to a) forest management approaches
to prevent large, high severity fires and b) management/restoration requirements after fires of
varying severities. Below we provide brief highlights for each module that received funding in
2011. The complete reports for each module follow the highlights.
Fuels and Fire Module
We investigated tree regeneration and shrub response within stand-replacing
patches for five recent fires. Natural tree regeneration was generally low across all
fires, particularly for pine species. Shrub cover was generally high, with over half
of the sampled patches exceeding 60% cover. Not surprisingly, areas that were
treated post-fire (salvage harvest with and without planting) resulted in noticeably
higher conifer regeneration.
We compared the impacts of the DFPZ network in Meadow Valley on modeled
fire behavior to the modeled impacts of a theoretical landscape fuel treatment
design based on an optimization algorithm. When averaged across the entire
landscape, burn probabilities for both treated scenarios (actual or theoretical) were
approximately half that for the untreated scenario in 2010, and were nearly
indistinguishable from each other. This suggests that the DFPZ network that was
implemented could be quite effective at reducing potential fire behavior.
Terrestrial Bird Module
We focused on monitoring birds in three burned areas. Sixteen of 36 avian species
investigated were significantly more abundant in burned areas and 14 were
significantly more abundant in unburned habitat in 2011. Avian community
metrics were significantly higher in unburned forest in 2011 than any of the three
burned areas. Among the three burned areas, the Storrie fire area had the highest
avian community indices. Avian community metrics were significantly lower in
both the Moonlight and Storrie fire private lands than in the adjoining Forest
6
Service lands. Burn severity was a significant predictor of avian diversity and
total bird abundance across burned areas with values highest at moderate burn
severity. We present 24 recommendations for managing burned areas for birds in
the context of a balanced approach to management of the Sierra Nevada
ecosystem.
California Spotted Owl Module
We completed our fourth year of monitoring CSO distribution and abundance in
the Meadow Valley Project Area. During 2008-2010 we observed similar
numbers of territorial owl sites (6-7) in the post-treatment landscape, although we
did observe territory-specific responses in both areas that have experienced
treatments and sites where little treatment has occurred. In 2011 we documented 4
territories within the Meadow Valley Project Area. This was a reduction of 3
territorial sites compared to 2010, whereas we had a net reduction of 2 additional
territorial sites across the rest of the PLS area
We completed analysis of the demographic data from the 20-year long
demographic study of CSO on the Lassen Demographic Study Area embedded
within our overall study. These results suggest that CSOs on the study area have
experienced a long-term population decline, with an estimated annual population
growth rate (lambda) of 0.986 (SE = 0.010) with 95% confidence intervals
ranging of 0.967-1.006. Our biggest challenge to date has been the lack of
accurate vegetation information to develop predictive habitat models and to
document changes in vegetation due to treatment or disturbance
Summary for 2011:
This work represents substantial scientific study that has occurred over the last ten years. Our
original expectation was to continue for up to another two years within the HFQLG Pilot Project
area to capture adequate post-treatment data. When we began this study the pilot project was
scheduled to end in 2005, and since then it has been extended twice, now to 2012, to enable the
complete pilot project to be implemented. Upon completion of the field work the remainder of
the effort will be devoted to data analysis and reporting. Any future work that is deemed
mutually desirable may continue (or start anew) as this phase of work winds down. That will
depend on the interest of forest managers and the availability of funding.
We understand there is some uncertainty and sometimes controversy over how various forest
elements will respond to planned forest management practices. This is likely to be the case under
any chosen management regime. The objective of PSW was to tackle the difficult scientific
challenges derived from the salient management questions. PSW, as a research organization,
remains wholly objective in executing this charge. We have assembled an excellent team of
scientists with the appropriate areas of expertise, and we have done the best we can to design our
work to address the important questions. Many of these questions present significant challenges
to experimental design of field ecology experiments, and management constraints further
constrain our ability to test questions with traditional hypothesis-testing approaches. We expect
7
to make the most of these opportunities in advancing our scientific understanding of forest
ecosystem response to management practices.
8
CHAPTER 1:
Fuels and Fire at the Landscape Scale
Dr. Brandon Collins, Christopher Dow, Gary Roller, Dr. Scott Stephens
Executive Summary
We present results for two different projects that have been conducted as part of the Storrie Fire
Restoration effort and the Plumas-Lassen Administrative Study. Brief descriptions and bullet
points are presented below each project title.
1. TREE REGENERATION AND SHRUB DYNAMICS IN STAND-REPLACING FIRE PATCHES
We quantified tree regeneration patterns and shrub dynamics for stand-replacing patches
within the five fairly recent fires that occurred within the Plumas National Forest.
• Although tree regeneration densities varied considerably across the five sampled
fires, over 50% of the sampled stand-replacing patches and approximately 80% all
plots had no tree regeneration.
• The percentage of patches and plots without pine regeneration was even higher, 72%
and 87%, respectively.
• Not surprisingly, plots that were salvage-harvested and then planted had the highest
pine densities, while plots that were salvage-harvested alone or were not treated had
for the most part zero pine regeneration, with the exception of a few outliers.
• The highest hardwood densities were associated with plots further from patch edges,
while the lowest densities were associated with plots closer to patch edges that were
on flatter slopes.
• The highest non-pine conifer regeneration density was associated with plots that were
salvaged-logged alone and occurred at elevations below 1635 m, as well as plots that
were on more north-facing aspects.
2. COMPARISON OF AN ACTUAL FUEL TREATMENT NETWORK TO A THEORETICAL TREATMENT
DESIGN IN MEADOW VALLEY
We compared the impacts of an actual landscape fuel treatment project on modeled fire
behavior to the modeled impacts of a landscape fuel treatment design based on an
optimization algorithm.
• When averaged across the entire landscape, conditional burn probabilities for both
treated scenarios (actual or theoretical) were approximately half that for the untreated
scenario in 2010.
• In 2010 and 2020 average burn probabilities for the two treatment scenarios were
nearly indistinguishable, but by 2030 average burn probability for the actual treatment
scenario increased slightly while the theoretical treatment scenario remained stable.
9
1. TREE REGENERATION AND SHRUB DYNAMICS IN STANDREPLACING FIRE PATCHES
Introduction:
Stand-replacing fire in forest types historically associated with low- to moderate-severity
fire has received considerable attention in recent years. Characterizations of fire patterns in these
forest types prior to Euro-American settlement vary in inferring the role of stand-replacing fire
historically, with some studies demonstrating that it was a relatively minor component of the fire
regime, accounting for a small proportion of the total area burned (Brown and Schoettle 2008;
Scholl and Taylor 2010), and other studies demonstrating stand-replacing fire did occur at larger
spatial extents, but occurred infrequently (Brown et al. 1999; Beaty and Taylor 2008).
Irrespective of these inconsistencies, it almost ubiquitously acknowledged that contemporary
patterns of stand-replacing fire in these forest types have deviated considerably from historical
patterns (Graham 2003; Keane et al. 2008; Miller et al. 2009; but see Williams and Baker 2012).
This is largely attributed to the changes in forest structure and composition associated with longestablished fire suppression and past land management practices (timber harvesting, livestock
grazing) (Parsons and Debenedetti 1979; Hessburg et al. 2005; Naficy et al. 2010; Collins et al.
2011). Another contributing factor could be that much of the area burned in contemporary
wildfires coincides with fairly extreme fire weather conditions, as these are the conditions under
which fires generally escape initial fire suppression efforts (Finney et al. 2011).
The impacts of this shift towards more extensive stand-replacing fire in forest types
historically associated with low- to moderate-severity fire are largely unclear. Some studies
suggest that the early-successional plant communities that develop following stand-replacing fire
contribute towards the maintenance of overall biodiversity (Odion et al. 2009; Swanson et al.
2011). However, given that many of the species associated with these forest types evolved
largely with low- to moderate-severity fire and can withstand or recover rapidly following such
fire (Agee 1993; Stephens et al. 2008), they generally have limited capacity to recover naturally
following extensive stand-replacing fire (Allen et al. 2002; Savage and Mast 2005). Previous
work from forests in the Klamath-Siskiyou Mountains, which have been historically associated
with mixed-severity fire, demonstrated highly variable but generally abundant conifer
regeneration in stand-replacing patches (Shatford et al. 2007; Donato et al. 2009). It is difficult to
know how applicable the results from these studies are in forest types historically associated with
low- to moderate-severity fire. Furthermore, it is unclear what proportion of conifer regeneration
is pine, which was a substantial component of these forest types historically and is critical for
achieving forest restoration objectives.
In this study we quantify tree regeneration patterns and shrub dynamics for standreplacing patches within the five recent fires that occurred in the northern Sierra Nevada. Our
initial intent was to sample patches that did not receive any post-fire management activities
(salvage-harvest, planting, etc.). However, field observations revealed evidence of post-fire
activities in some sampled patches, which were incorporated in the study. Our null hypotheses
were: 1) tree regeneration would be variable but on average relatively abundant in standreplacing patches (>100 trees/ha); 2) regeneration would be relatively even among species
groups (e.g., Pinus, non-Pinus conifers, hardwood); 3) Tree regeneration would be negatively
related to both shrub cover and distance-to-edge of patch.
10
Methods:
Study area and field sampling
Our study area is located in the Plumas National Forest, situated in the northern Sierra
Nevada at 39°56’N, 121°3’W (Figure 1). The study area was defined by five fairly recent fires
that burned in relatively close proximity to one another: Lookout (1999), Pigeon (1999), Bucks
Complex (1999), Storrie (2000), and Rich (2008) (Figure 1, Table 1). We used satellite-derived
estimates of fire severity to identify stand-replacing patches within these five fires. The fire
severity estimates were based on the relative differenced Normalized Burn Ratio (RdNBR),
which is computed from Landsat TM imagery (see Miller and Thode 2007 and Miller et al.
2009b for more specific explanation of methodologies). This index has been used extensively to
characterize relatively recent fires (Safford et al. 2008; Collins et al. 2009) and fire regimes
(Holden et al. 2007; van Wagtendonk and Lutz 2007; Miller et al. 2009a). Miller et al. (2009b)
report user’s and producer’s accuracies for RdNBR-based high severity classification to range
between 70.7% and 85.3%, indicating RdNBR robustly captures stand-replacing fire effects.
Using classified RdNBR images (threshold values established by Miller and Thode 2007)
we employed the patch delineation algorithm PatchMorph (Girvetz and Greco 2007) to identify
contiguous patches of stand-replacing fire that were > 2 ha. The PatchMorph tool can be used in
ArcGIS and has been used previously to delineate stand-replacing patches (Collins and Stephens
2010). We chose a minimum patch size of 2 ha because we were interested in characterizing tree
regeneration and shrub response within larger areas that were generally void of substantial tree
seed sources.
Our field sampling efforts were constrained by access, crew safety concerns, and time. As
such, we were only able to sample a portion of the patches identified by the PatchMorph tool
(Figure 1). Field plot locations were established using a systematic grid of points with 200 m
spacing, which were based on a random starting location for each stand-replacing patch. We
intensified sampling near patch edges to investigate potential distance-to-edge effects on tree and
shrub dynamics. The rules for plot intensification were as follows: if grid points were 1) <25 m
from patch edge, there were no additional plots; 2) 25-75 m from patch edge, a plot was added
that was 25 m towards the nearest point on the patch edge; 3) 75-125 m from patch edge, plots
were added at 50 m both towards and away from the nearest point on the patch edge; and 4)
>125 m from patch edge, there were no additional plots. We sampled 26 stand-replacing patches
across the five fires, with a total of 277 field plots. Five of the 26 sampled patches were treated
post-fire: three were salvage-logged only, and two were salvage-logged and planted. These
treated patches accounted for 112 field plots.
At each plot we established three 10-meter transects using a random starting azimuth for
the first transect and then adding 120 degrees for the next two (Figure 2). Along each transect we
measured shrub length, average shrub height, and ground cover. Ground cover was broken down
into cover classes of woody shrubs, herbaceous plants, grass, litter, wood (downed trees), rock,
and bare mineral soil. Tree regeneration was sampled on the same three transects, with trees
>0.1 cm diameter-at-breast-height tallied by species on a two meter wide belt and established
seedlings <1.4 m tall tallied by species in circular plots at the end of the transect (Figure 2).
Data analysis
We performed analysis at two levels, individual plots and stand-replacing patches, with
the intent of exploring both site- and patch-level influences on tree regeneration. For each plot
11
and sampled stand-replacing patch, we assembled a suite of potential explanatory variables that
characterize the pre- and post-fire vegetation, topography, and patch characteristics. Pre-fire
vegetation was based on a 1997 Landsat-derived map produced the U. S. Forest Service Region
5. The map provided broad vegetation type (mixed-conifer, mixed hardwood, mixed chaparral,
and white fir) and canopy closure (25-39%, 40-59%, and 60-100%) classes. Post-fire shrub cover
and average shrub height we based on transect data from individual plots. Topographic variables
included slope steepness, aspect, and elevation. For the plot-level analysis the topographic
variables were based on field observations, with elevation acquired via handheld GPS. For the
patch-level analysis these variables were derived from a 10 m digital elevation model. Additional
potential explanatory variables included post-fire management (salvage-logged only, salvagelogged and planted, or none), as well as distance-to-patch-edge at the individual plot level and
patch area and perimeter-to-area ratio at the patch level.
We used these potential explanatory variables to explore possible patterns in tree
regeneration densities, which we binned in three groups: pines (Pinus ponderosa and P.
lambertiana), non-pine conifers (Abies concolor, A. magnifica, Calocedrus decurrens, and
Pseudotsuga menziesii), and hardwood (Quercus kelloggii). At the plot level, tree tallies were
multiplied by the appropriate scaling factor based the tree size/sampling area to obtain per
hectare tree regeneration densities (Figure 2). Patch-level densities were based on the average of
all plots within a given patch. We explored possible relationships between these three tree
regeneration groups and the potential explanatory variables using regression tree analysis.
Regression tree analysis offers clear advantages over traditional linear models because it can
handle nonlinear or discontinuous relationships between variables and high-order interactions
(De'ath and Fabricius 2000). In addition, the hierarchical structure and identification of potential
threshold values for independent variables is well suited for explaining ecological phenomena
(De'ath and Fabricius 2000; Collins et al. 2009). The regression tree is constructed by repeatedly
splitting the data into increasingly homogenous groups based on the dependent variable. Each
split minimizes the sum of squares within the resulting groups. The number of terminal nodes
was determined using the one-standard error rule on the cross-validated relative error (Breiman
et al. 1984; De'ath 2002). We ran multiple iterations using this method to confirm the chosen
number of terminal nodes. Tree regeneration densities were log-transformed due the very large
range in observed densities (Figure 3).
Results:
Although tree regeneration densities varied considerably across the five sampled fires,
over 50% of the sampled stand-replacing patches and approximately 80% all plots had no tree
regeneration (Figure 3). The percentage of patches, and to a lesser extent plots, without pine
regeneration was even higher, 72% and 87%, respectively. However, nearly 20% of the sampled
patches had non-pine conifer regeneration in excess of 2000 trees/ha (Figure 3). Of the three
species groups, hardwood was clearly the most common at intermediate densities, with almost
30% of the sampled patches averaging between 400 and1400 trees/ha. Shrub cover was generally
high, with approximately 60% of both the patches and individual plots exceeding 60% cover
(Figure 4). Average shrub height across all plots with shrubs present was 86 cm. Although there
were some plots with no shrub cover (n = 14), all patches on average had shrub cover > 0 (Figure
4).
12
The regression tree analysis at the patch level identified shrub cover as the only
significant predictor of both non-pine conifer and hardwood regeneration density (results not
shown). For both species groups average regeneration density tended to be higher in patches with
average shrub cover over 25%. The same analysis failed to identify any significant variables
explaining the variability in pine regeneration density. At the plot-level the most important factor
explaining non-pine conifer regeneration was post-fire management (Figure 5). The next most
important explanatory factors were elevation and aspect (Figure 5). The highest non-pine conifer
regeneration density was associated with plots that were salvaged-logged alone and occurred at
elevations below 1635 m, as well as plots that were on more north-facing aspects. The lowest
non-pine conifer regeneration density was associated with plots that were salvaged-logged alone
and occurred at elevations above 1635 m (Figure 5). The significant factors explaining the
variability in hardwood regeneration density were distance to nearest patch edge, slope, and
shrub height. The highest hardwood densities were associated with plots further from patch edge,
while the lowest densities were associated with plots closer to patch edges that were on flatter
slopes (Figure 5). Post-fire management activity was the only variable that significantly
explained variability in plot-level pine regeneration density. Not surprisingly, plots that were
salvage-harvested and then planted had the highest pine densities, while plots that were salvageharvested alone or were not treated had for the most part zero pine regeneration, with the
exception of a few outliers (Figure 5).
13
2. COMPARISON OF AN ACTUAL FUEL TREATMENT NETWORK TO A
THEORETICAL TREATMENT DESIGN IN MEADOW VALLEY
Introduction:
Fire exclusion policies coupled with past management practices have rendered many
western U. S. forests susceptible to uncharacteristically large and intense wildfires (Hessburg et
al. 2005; Stephens et al. 2009). Large-scale fuel reduction efforts are needed in many forests to
mitigate the potential for extensive losses of productive and mature forests from such wildfires.
However, there are a number of land management obligations and operational constraints that
limit the extent that fuel reduction treatments can be applied across many landscapes (Collins et
al. 2010). As a result, forest managers are tasked with devising strategic and innovative
treatments across a landscape to achieve multiple objectives, including mitigating the potential
for extensive fire-caused tree mortality throughout the area. Recent simulation studies have
demonstrated reductions in both potential fire behavior (Finney et al. 2007; Collins et al. 2011)
and expected tree mortality (Ager et al. 2007; Ager et al. 2010a) with as little as 10-25% of the
landscape treated when employing strategic treatments. While there have been significant
advancements in both the tools and approaches available to guide effective landscape fuel
treatment design (Ager et al. 2006; Finney 2007), few examples exist “on the ground” (but see
Moghaddas et al. 2010). This is likely a product of the complexities associated with not only
planning across large landscapes, but acquiring necessary data and executing models as well
(Collins et al. 2010).
Theoretical design of landscape fuel treatment projects has evolved from an arrangement
of regularly spaced, staggered treatment units (SPLATS - Finney 2001) to a more spatially
explicit arrangement informed by an optimization algorithm called the Treatment Optimization
Model (TOM - Finney 2004, 2006, 2007). TOM identifies the “ideal” areas to treat, such that fire
spread across a landscape is minimized based on user-identified potential treatment areas. This
allows for the incorporation of real-world constraints on fuel treatment activities (e.g., sensitive
habitat, stream buffers, and inaccessible areas) and has the potential to yield a more realistic
treatment arrangement than a design based on regular treatment arrangements. As an alternative
to the intensive modeling procedures associated with TOM, there are examples of recent
landscape fuel treatment projects in which the treatment arrangement was based largely on local
knowledge and intuition (Meadow Valley Project, Plumas National Forest; Last Chance Project,
Tahoe National Forest). Recent studies have demonstrated that these projects are quite effective
at reducing modeled landscape-level fire behavior (Moghaddas et al. 2010; Collins et al. 2011).
It is unknown how the reduction in modeled fire behavior associated with these projects
compares to a theoretical treatment design informed by an optimization algorithm. Our intent
with the present study was to explicitly address this: i.e., compare the impacts an actual
landscape fuel treatment project to a design based on TOM which incorporates some of the
actual constraints on the landscape.
Methods:
Study area
Our study area is located in the Plumas National Forest, situated in the northern Sierra
Nevada at 39°56’N, 121°3’W (Figure 6). The study area boundary is defined by three adjacent
14
Hydrologic Unit Code sixth-level watersheds, with a slight modification to the southern-most
watershed based on the extent of the Meadow Valley Project area (USDA 2004a). The climate is
Mediterranean with a predominance of winter precipitation totaling about 1050 mm per year
(Ansley and Battles 1998). The core study area is 19,236 ha, with elevations ranging from 8502100 m (Figure 6). Vegetation on this landscape is primarily mixed conifer forest (Schoenherr
1992; Barbour and Major 1995), consisting of white fir (Abies concolor), Douglas-fir
(Pseudotsuga menziesii), sugar pine (Pinus lambertiana), ponderosa pine (P. ponderosa), Jeffrey
pine (P. jeffreyi), incense-cedar (Calocedrus decurrens), California black oak (Quercus
kelloggii), and other less common hardwood species. Red fir (A. magnifica) is common at higher
elevations, where it mixes with white fir. In addition, a number of species are found occasionally
in or on the edge of the mixed conifer forest: western white pine (P. monticola) at higher
elevations, lodgepole pine (P. contorta) in cold air pockets and riparian zones, western juniper
(Juniperus occidentalis) on dry sites, and California hazelnut (Corylus cornuta), dogwood
(Cornus spp.) and willow (Salix spp.) in moister sites. Montane chaparral and some meadows are
interspersed with the forest. Tree density varies as a result of recent fire and timber management
history, elevation, slope, aspect, and edaphic conditions. Historical fire occurrence, inferred from
fire scars recorded in tree rings, suggests a fire regime with predominantly frequent, low- to
moderate-severity fires occurring at intervals ranging from 7 to 12 years (Moody et al. 2006).
The projects that contributed to the fuel treatment network in our study area are part of
the larger Herger-Feinstein Quincy Library Group Pilot Project. This project was directed by the
U.S. Congress to enable the implementation of a local community forest management vision that
attempts to address a combination of forest management objectives including forest health,
reduction of high severity fire, conservation of wildlife habitats, and stabilization of economic
conditions in the local community. The projects in Meadow Valley encompassed a range of
treatment types and intensities. This range in treatment types and intensities is a product of both
changes in regional management directions (USDA 2001, 2004b) and differing land management
constraints across a complex landscape (Collins et al. 2010; Moghaddas et al. 2010). These
treatments collectively covered 3692 ha, or 19% of our study area (Figure 6), and were
implemented between 2003 and 2008.
Field data collection
We used data from two distinct field sampling efforts. The first effort was aimed at a
landscape-scale characterization using a stratified-random approach to establish plots. This
approach, referred to hereafter as landscape plots, included four strata: slope (3 levels: <15%, 1530%, >30%), elevation (3 levels: <1400 m, 1400-1600 m, and >1600 m), aspect (4 levels: N, E,
S, and W), and dominant vegetation (6 levels based on air photo-interpreted California Wildlife
Habitat Relationship vegetation classes) (VESTRA 2003). Plot locations were randomly
assigned within each strata combination. In total there were 604 landscape plots, which were
sampled between 2004 and 2006. The second field sampling effort was focused on treated areas,
referred to hereafter as treatment plots. Plot locations were selected to capture the range in both
treatment types and geographic locations of treatments throughout the study area. There were 72
treatment plots. These plots were initially sampled 1-3 years prior to treatments, and re-sampled
1 year following treatments. Pre-treatment sampling was conducted between 2002 and 2007,
with post-treatment re-sampling between 2004 and 2009.
The two sampling efforts followed different plot sampling methodology based on the
different objectives of the two efforts. Landscape plots were smaller and less intensive than the
15
treatment plots to allow for a greater number of plots that covered a much larger spatial extent.
Landscape plots were circular with a fixed radius of 12.6 m, resulting in a plot area of 0.05 ha.
Trees >1.37 m tall and <10cm diameter-at-breast-height (dbh) were tallied by species with no
individual height measurements. Trees ≥10 cm dbh were binned into 10 cm dbh classes with
their heights measured to the nearest 1m for trees <10 m, while heights for trees ≥10 m were
estimated to nearest 10 m category (e.g., 10-20 m, 20-30 m). Crown base height for all trees
≥10 cm dbh was measured to the nearest 1 m. Within each plot downed woody fuels, litter, duff,
and shrubs were sampled on two 12 m transects, with random directions radiating from plot
center. Downed woody fuels were sampled using the planar intercept method (Brown 1974) with
different lengths of the transect for different fuel particle size classes: 1-h (0-0.64 cm) and 10-h
(0.65-2.54 cm) fuels were sampled from 10 to 12 m, 100-h (2.55-7.62 cm) fuels from 9 to 12 m,
and 1000-h fuels from 2 to 12 m with diameters and sound/rotten recorded for each piece. Litter
and duff depths and fuel height was recorded at 5 m and 12 m. Shrub cover was measured as
intersections along each transect with species and average height recorded for each individual.
Treatment plots were rectangular: 50 m X 20 m, resulting in a plot area of 0.1 ha. Tree sampling
was as follows: ≥76.2 cm dbh sampled on the entire plot, 40.6 to 76.2 cm dbh on the center half
of the plot, 12.7 to 40.5 cm dbh on the center quarter of the plot, 2.5 to 12.7 cm dbh on five
16 m2 sub-plots established along the main plot centerline. Total height, crown base height, and
dbh were measured on all trees. Shrub cover by species was recorded at each of the five 16 m2
sub-plots.
Data integration
We used tree data from both field sampling efforts to generate tree lists to ‘populate’
distinct forested stands across the landscape. Stand boundaries corresponded with mapped
vegetation polygons (VESTRA 2003). There were 1505 forested stands that were either within or
intersected our core study area boundary, ranging in size from 0.5 to 490 ha. Forested stands
accounted for approximately 92% of the core study area. Non-forest vegetation, barren areas, and
lakes accounted for the remaining 8% of the study area. Tree lists from plots were assigned to
forested stands using an imputation approach that was based on the same four strata that were
used in sampling landscape plots: slope, elevation, aspect, dominant vegetation type. The rules
for imputing went as follows: if multiple plots matched the four strata for a given stand, a
random plot was chosen; if no plots matched exactly, strata were eliminated in the following
order until a match was made: slope, then aspect, then elevation.
Polygons for the actual treatments were compiled by the Herger-Feinstein Quincy
Library Group monitoring team (http://www.fs.fed.us/r5/hfqlg/monitoring/) and were overlaid on
our stand/vegetation polygon layer. In cases where stands were partially overlapped by treatment
polygons, stands were sub-divided such that stand boundaries corresponded with treatment
boundaries. Treatment plots were assigned to stands within treatment polygons based on the five
treatment types described in Table 2. As with our other plot imputation approach when multiple
plots existed for a given treatment type, a random plot was chosen for a given treated stand. To
ensure consistency between the pre- and post-treatment landscape conditions, plot assignments
for treated stands superseded strata-based assignments mentioned previously. In the specific
cases where a treatment plot fell in a given treated stand that plot was assigned to the stand.
Polygons for the theoretical treatments were derived from the TOM module within the
spatial fire behavior model FlamMap (Finney 2006). In order to execute TOM users need to
create an “ideal landscape,” which identifies the vegetation and fuel conditions for all areas
16
where treatments are possible. To identify possible treatment stands within our study area we
assembled spatial layers of land allocations (Moghaddas et al. 2010), which were provided by the
Plumas National Forest staff . The particular land allocations where treatments were not possible
included: Protected Activity Centers (spotted owl and northern goshawk), Spotted Owl Habitat
Areas, Offbase and Deferred areas, and Stream buffers. The removal of these land allocations
from the potentially treatable land base is consistent with recent Forest Service fuel treatment
projects throughout the Sierra Nevada. We used the range of sampled vegetation and fuel
conditions from the treatment plots (post-treatment) to define the conditions within treatable
stands for the “ideal landscape.” TOM identifies optimal treatment areas by contrasting fire
behavior predictions between the “ideal landscape” and an untreated landscape with the intent of
disrupting fire spread (Finney 2006). We limited the identified treatment areas, which in some
cases were smaller than 0.25 ha, to only those 10 ha in order to better reflect the reality of
potential implementation.
We created three databases with tree lists for every forested stand across our buffered
study area : 1) an untreated database, in which stands that fell within treatment boundaries were
‘populated’ by tree lists from pre-treatment sampling efforts for the treatment plots and stands
outside treatment polygons were ‘populated’ using the imputation approach described
previously; 2) a treated database based on the actual treatments, in which tree lists for treated
stands were generated from the post-treatment plot measurements of the same plots used in the
pre-treatment database, while tree lists for stands outside of treatment polygons were the same as
those in the pre-treatment database; and 3) a treated database based on theoretical treatment
stands that were identified based on outputs from TOM (Finney 2006). Tree lists for TOM stands
were based on the same post-treatment plot measurements as in 2). The three tree list databases
were entered into the Forest Vegetation Simulator (FVS)(Dixon 2002) to model forest dynamics
within the treated and untreated landscapes for three 10-yr cycles beyond our starting year of
2010. This modeling was performed using the integrated platform ArcFuels (Ager et al. 2006;
Ager et al. 2011), which executes FVS and the fire and fuels extension (FFE) to generate the
necessary stand structure inputs needed to calculate key fire behavior parameters from runs of
spatially explicit fire behavior models: canopy bulk density, canopy cover, canopy height, and
canopy base height.
Fire modeling
We obtained weather information from the Quincy and Cashman Remote Automated
Weather Stations (RAWS), restricting the analysis period to the dominant fire season for the area
(June 1 – September 30) (Figure 6). Observations were available from 2002 to 2009. We used
90th percentile and above wind speeds, based on hourly observations, to generate multiple wind
scenarios under which fires were simulated. We identified the dominant direction and average
speed of all observations at or above the 90th percentile value, 24 km hr-1. This resulted in four
different dominant wind directions, each with its own wind speed and relative frequency (based
on the proportion of observations recorded at or above the 90th percentile value for each
dominant direction) (Table 2). The modeled wind speeds were similar to those recorded during
large spread events in the nearby 2008 Rich fire. We used 97th percentile fuel moistures, as these
are the conditions associated with large fire growth and difficulty in control.
We derived the necessary topographic inputs – slope, aspect, and elevation – using a
10 m digital elevation model obtained from the National Elevation Dataset (http://ned.usgs.gov/).
Stand structure and fuels layers were derived from FVS outputs. For each stand, under each
17
scenario, FVS outputs for canopy cover, canopy bulk density, canopy base height, and dominant
tree height, along with a fuel model assignment (computed outside of FVS), were compiled to
develop continuous layers for each of these five variables across the buffered Meadow Valley
study area (Figure 6). The buffered area included a 2 km buffer around the core study area. To
avoid potential edge effects we extracted the RANDIG output (see below) from only the core
Meadow Valley study area, i.e., not including the area within the 2 km buffer (Figure 6).
We employed a command-line version of FlamMap (Finney 2006) called RANDIG to
model fires across the Meadow Valley landscape (Figure 6). RANDIG uses the minimum travel
time method (Finney 2002) to simulate fire spread based on user-inputs for number/pattern of
ignitions, fire duration, wind speed and direction, fuel moistures, topography, stand structure,
and fuels. For each scenario (untreated, treated-ACTUAL, treated-THEORETICAL) and time step
(2010, 2020, 2030), we simulated 10,000 randomly placed ignitions, burning for 240 minutes
(one 4-hour burn period). This burn period duration was selected such that simulated fire sizes
(for one burn period) approximated large-spread events (daily) observed in a nearby recent
wildfires: Wheeler (2007), Moonlight (2007), and Rich (2008) (Ager et al. 2010a). Large daily
spread events in these fires ranged between 1000 and 6000 ha (Fites et al. 2007; Dailey et al.
2008), which is consistent with the interquartile range (25th to 75th percentile) throughout the
simulation duration for the untreated landscape.
For each scenario and time step, RANDIG outputs overall conditional burn probabilities,
which are computed by dividing the total number of times a pixel burned by the total number of
simulated fires (n = 10,000), as well as marginal conditional burn probabilities for 20 flame
length classes (0-10 m in 0.5 m increments). To separate out more problematic simulated fire
occurrence, both from a fire effects and a fire suppression standpoint, we only performed
analysis on the burn probabilities where modeled flame lengths were greater than 2 m, i.e., we
summed marginal burn probabilities for all flame length classes >2 m. This threshold has been
used to identify more problematic fire behavior in previous work (Collins et al. 2011; Ager et al.
2012). We imported the burn probabilities into ArcGIS software for further data analysis. For
each of the three scenarios at each time step we computed a mean conditional burn probability
(flame lengths >2 m), using only those pixels within the core study area (Figure 6).
Results:
Across much of the Meadow Valley landscape the conditional probabilities of fire
occurring with flame lengths >2 m for the initial untreated scenario (2010) noticeably exceeded
those for either treated scenario (Figure 7). The central-eastern portion of the landscape had the
highest burn probabilities for the untreated scenario. Burn probabilities in this area were reduced
considerably under the treated-ACTUAL scenario, but still remained high relative to the rest of the
landscape. The treated-THEORETICAL scenario was more effective than the treated- ACTUAL
scenario at reducing burn probabilities in this central-eastern portion of the landscape, but had
slightly higher burn probabilities in the southern portion of the landscape (Figure 7). Both treated
scenarios effectively reduced burn probabilities in the north-eastern portion of the landscape
(Figure 7).
When averaged across the entire landscape, conditional burn probabilities for both treated
scenarios were approximately half that for the untreated scenario in 2010 (Figure 8). Burn
probabilities for all three scenarios declined in 2020, with the untreated scenario declining more
substantially (Figure 8). In 2010 and 2020 average burn probabilities for the two treatment
18
scenarios were nearly indistinguishable, but by 2030 average burn probability for the treatedACTUAL scenario increased slightly while the treated-THEORETICAL scenario remained stable.
Average burn probability for the untreated scenario was unchanged between 2020 and 2030
(Figure 8).
Our findings demonstrating that coordinated fuel treatments across a landscape can
substantially reduce potential fire behavior are not novel; several previous studies have
demonstrated this (Ager et al. 2007; Finney et al. 2007; Ager et al. 2010b; Moghaddas et al.
2010; Collins et al. 2011). What is more novel about our study is the comparison of two
landscape fuel treatment designs, one that has actually been implemented and one that is
theoretical. The fact that modeled burn probabilities were considerably reduced in both treatment
scenarios relative to the untreated scenario suggests that both the actual and theoretical treatment
designs can be quite effective at reducing more problematic fire behavior. While both designs
were effective, our findings indicate that the theoretical design resulted in a more even reduction
of burn probabilities across the Meadow Valley landscape (Figure 7). This is not too surprising
given that the theoretical design is informed by an algorithm that seeks to disrupt major flow
paths for fire (Finney 2006), while the actual treatment design was based more on intuition and
circumstance. The more surprising fact is that there was not more of a gap between the actual
design and the theoretical design. This is especially true given how different the topologies were
between the two designs (Figure 6). We expected that the treatment areas identified by the
optimization algorithm would have resulted in a much more effective reduction of modeled
problematic fire, which was not the case (Figure 8).
One of the factors driving the similarity in overall average results between the two
treatment scenarios may be that there is a considerable amount of area in the Meadow Valley
landscape that was removed from the potentially treatable land base due to the various
constraints imposed by the different land designations. The various constraints (wildlife habitat,
riparian buffers, and other reserves) accounted for 24% of the core study area. Finney et al.
(2007) point out that as the proportion of non-treatable area approaches 40% for a given
landscape the optimizing routine is essentially no different from random treatment placement.
References:
Agee JK 1993. Fire ecology of Pacific Northwest forests. Island Press, Washington D.C., USA.
Ager AA, Bahro B and Barber K 2006. Automating the Fireshed Assessment Process with
ArcGIS. In Fuels management - how to measure success, pp. 163-168. U. S. Department
of Agriculture, Forest Service, Rocky Mountain Research Station, Portland, OR.
Ager AA, Finney MA, Kerns BK and Maffei H 2007. Modeling wildfire risk to northern spotted
owl (Strix occidentalis caurina) habitat in Central Oregon, USA. Forest Ecol Manage
246: 45-56.
Ager AA, Vaillant NM and Finney MA 2010a. A comparison of landscape fuel treatment
strategies to mitigate wildland fire risk in the urban interface and preserve old forest
structure. Forest Ecol Manage 259: 1556-1570.
Ager AA, Finney MA, McMahan A and Cathcart J 2010b. Measuring the effect of fuel
treatments on forest carbon using landscape risk analysis. Nat Hazards Earth Syst Sci 10:
2515-2526.
Ager AA, Vaillant NM, Anderson J and Miller L 2011. ArcFuels User Guide: for use with
ArcGIS 9.X. Internal Report. US Department of Agriculture, Forest Service, Pacific
19
Northwest Research Station, Western Wildland Environmental Threat Assessment
Center, Prineville, OR, USA. p. 245.
Ager AA, Vaillant NM, Owens DE, Brittain S and Hamann J 2012. Overview and example
application of the Landscape Treatment Designer. General Technical Report PNW-GTR859. U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station,
Portland, OR, USA. p. 11.
Allen CD, Savage M, Falk DA, Suckling KF, Swetnam TW, Schulke T, Stacey PB, Morgan P,
Hoffman M and Klingel JT 2002. Ecological restoration of southwestern ponderosa pine
ecosystems: a broad perspective. Ecol Appl 12: 1418-1433.
Barbour MG and Major J (eds.) 1995. Terrestrial Vegetation of California: New Expanded
Edition. California Native Plant Society, Davis.
Beaty RM and Taylor AH 2008. Fire history and the structure and dynamics of a mixed conifer
forest landscape in the northern Sierra Nevada, Lake Tahoe Basin, California, USA.
Forest Ecol Manage 255: 707-719.
Breiman L, Friedman JH, Olshen RA and Stone CG 1984. Classification and regression trees.
Wadsworth, Belmont, CA, USA.
Brown JK 1974. Handbook for inventorying downed woody material. General Technical Report
INT-16. U.S. Department of Agriculture, Forest Service, Intermountain Forest and Range
Experiment Station, Odgen, UT. p. 32.
Brown PM, Kaufmann MR and Shepperd WD 1999. Long-term, landscape patterns of past fire
events in a montane ponderosa pine forest of central Colorado. Landscape Ecol 14: 513532.
Brown PM and Schoettle AW 2008. Fire and stand history in two limber pine (Pinus flexilis) and
Rocky Mountain bristlecone pine (Pinus aristata) stands in Colorado. Int J Wildland Fire
17: 339-347.
Collins BM, Miller JD, Thode AE, Kelly M, van Wagtendonk JW and Stephens SL 2009.
Interactions among wildland fires in a long-established Sierra Nevada natural fire area.
Ecosystems 12: 114-128.
Collins BM and Stephens SL 2010. Stand-replacing patches within a ‘mixed severity’ fire
regime: quantitative characterization using recent fires in a long-established natural fire
area. Landscape Ecol 25: 927-939.
Collins BM, Stephens SL, Moghaddas JJ and Battles J 2010. Challenges and approaches in
planning fuel treatments across fire-excluded forested landscapes. J Forest 108: 24-31.
Collins BM, Everett RG and Stephens SL 2011a. Impacts of fire exclusion and managed fire on
forest structure in an old growth Sierra Nevada mixed-conifer forest. Ecosphere 2: art 51.
Collins BM, Stephens SL, Roller GB and Battles JJ 2011b. Simulating fire and forest dynamics
for a landscape fuel treatment project in the Sierra Nevada. Forest Sci 57: 77-88.
Dailey S, Fites J, Reiner A and Mori S 2008. Fire behavior and effects in fuel treatments and
protected habitat on the Moonlight fireAvailable online
http://www.fs.fed.us/adaptivemanagement/projects/FBAT/docs/MoonlightFinal_8_6_08.
pdf, last accessed Apr. 8, 2010. Fire Behavior Assessment Team.
De'ath G and Fabricius KE 2000. Classification and regression trees: a powerful yet simple
technique for ecological data analysis. Ecology 81: 3178-3192.
De'ath G 2002. Multivariate regression trees: a new technique for modeling species-environment
relationships. Ecology 83: 1105-1117.
20
Dixon GE 2002. Essential FVS: a user's guide to the Forest Vegetation Simulator. Internal
Report. USDA, Forest Service, Rocky Mountain Research Station, Fort Collins, CO,
USA. p. 209.
Donato DC, Fontaine JB, Campbell JL, Robinson WD, Kauffman JB and Law BE 2009. Conifer
regeneration in stand-replacement portions of a large mixed-severity wildfire in the
Klamath-Siskiyou Mountains. Can J For Res 39: 823-838.
Finney MA 2001. Design of regular landscape fuel treatment patterns for modifying fire growth
and behavior. Forest Sci 47: 219-228.
Finney MA 2002. Fire growth using minimum travel time methods. Can J For Res 32: 14201424.
Finney MA 2004. Landscape fire simulation and fuel treatment optimization. In Hayes JL, Ager
AA and Barbour JR (eds.), Methods for integrated modeling of landscape change, pp.
117-131. USDA Forest Service, Pacific Northwest Research Station, Portland, OR, USA.
Finney MA 2006. An overview of FlamMap modeling capabilities. In Fuels management - how
to measure success, pp. 213-220. U. S. Department of Agriculture, Forest Service, Rocky
Mountain Research Station, Portland, OR.
Finney MA 2007. A computational method for optimising fuel treatment locations. Int J
Wildland Fire 16: 702-711.
Finney MA, Seli RC, McHugh CW, Ager AA, Bahro B and Agee JK 2007. Simulation of longterm landscape-level fuel treatment effects on large wildfires. Int J Wildland Fire 16:
712-727.
Finney MA, McHugh CW, Grenfell IC, Riley KL and Short KC 2011. A simulation of
probabilistic wildfire risk components for the continental United States. Stoch Environ
Res Risk Assess 25: 973-1000.
Fites J, Campbell M, Reiner A and Decker T 2007. Fire behavior and effects related to
suppression, fuel treatments, and protected areas on the Antelope Complex, Wheeler
fireAvailable online
http://www.fs.fed.us/adaptivemanagement/projects/FBAT/docs/Antelope_FINAL3_12_0
4_07.pdf; last accessed Apr. 8, 2010. Fire Behavior Assessment Team.
Girvetz EH and Greco SE 2007. How to define a patch: a spatial model for hierarchically
delineating organism-specific habitat patches. Landscape Ecol 22: 1131-1142.
Hessburg PF, Agee JK and Franklin JF 2005. Dry forests and wildland fires of the inland
Northwest USA: contrasting the landscape ecology of the pre-settlement and modern
eras. Forest Ecol Manage 211: 117-139.
Holden ZA, Morgan P, Crimmins MA, Steinhorst RK and Smith AM 2007. Fire season
precipitation variability influences fire extent and severity in a large southwestern
wilderness area, United States. Geophys Res Let
34: L16708.
Miller JD and Thode AE 2007. Quantifying burn severity in a heterogeneous landscape with a
relative version of the delta Normalized Burn Ratio (dNBR). Remote Sens Environ 109:
66-80.
Miller JD, Safford HD, Crimmins M and Thode AE 2009a. Quantitative evidence for increasing
forest fire severity in the Sierra Nevada and southern Cascade Mountains, California and
Nevada, USA. Ecosystems 12: 16-32.
Miller JD, Knapp EE, Key CH, Skinner CN, Isbell CJ, Creasy RM and Sherlock JW 2009b.
Calibration and validation of the relative differenced Normalized Burn Ratio (RdNBR) to
21
three measures of fire severity in the Sierra Nevada and Klamath Mountains, California,
USA. Remote Sens Environ 113: 645-646.
Moghaddas JJ, Collins BM, Menning K, Moghaddas EEY and Stephens SL 2010. Fuel treatment
effects on modeled landscape level fire behavior in the northern Sierra Nevada. Can J For
Res 40: 1751-1765.
Moody TJ, Fites-Kaufman J and Stephens SL 2006. Fire history and climate influences from
forests in the northern Sierra Nevada, USA. Fire Ecol 2: 115-141.
Naficy C, Sala A, Keeling EG, Graham J and DeLuca TH 2010. Interactive effects of historical
logging and fire exclusion on ponderosa pine forest structure in the northern Rockies.
Ecol Appl 20: 1851-1864.
Odion DC, Moritz MA and DellaSala DA 2009. Alternative community states maintained by fire
in the Klamath Mountains, USA. J Ecol: doi: 10.1111/j.1365-2745.2009.01597.x.
Parsons DJ and Debenedetti SH 1979. Impact of fire suppression on a mixed-conifer forest.
Forest Ecol Manage 2: 21-33.
Safford HD, Miller JD, Schmidt D, Roath B and Parsons A 2008. BAER soil burn severity maps
do not measure fire effects to vegetation: a comment on Odion and Hanson (2006).
Ecosystems 11: 1-11.
Savage M and Mast JN 2005. How resilient are southwestern ponderosa pine forests after crown
fires? Can J For Res 35: 967-977.
Schoenherr AA 1992. A Natural History of California. University of California Press, Berkeley.
Scholl AE and Taylor AH 2010. Fire regimes, forest change, and self-organization in an oldgrowth mixed-conifer forest, Yosemite National Park, USA. Ecol Appl 20: 362-380.
Shatford JPA, Hibbs DE and Puettmann KJ 2007. Conifer regeneration after forest fire in the
Klamath-Siskiyous: how much, how soon? J Forest 105: 139-146.
Stephens SL, Fry D and Franco-Vizcaíno E 2008. Wildfire and forests in northwestern Mexico:
the United States wishes it had similar fire problems. Ecol Soc 13: 10.
Stephens SL, Moghaddas JJ, Ediminster C, Fiedler CE, Hasse S, Harrington M, Keeley JE,
Knapp EE, McIver JD, Metlen K, Skinner CN and Youngblood A 2009. Fire treatment
effects on vegetation structure, fuels, and potential fire severity in western U.S. forests.
Ecol Appl 19: 305-320.
Swanson ME, Franklin JF, Beschta RL, Crisafulli CM, DellaSala DA, Hutto RL, Lindenmayer
DB and Swanson FJ 2011. The forgotten stage of forest succession: early-successional
ecosystems on forest sites. Front Ecol Environ 9: 117-125.
USDA 2001. Sierra Nevada forest plan amendment. Final Supplemental Environmental Impact
Statement. USDA, Forest Service, Pacific Southwest Region.
USDA 2004a. Sierra Nevada forest plan amendment. Final Supplemental Environmental Impact
Statement R5-MB-046. USDA, Forest Service, Pacific Southwest Region.
USDA 2004b. Environmental Assessment: Meadow Valley Defensible Fuel Profile Zone and
Group Selection Project. 113p.
van Wagtendonk JW and Lutz JA 2007. Fire regime attributes of wildland fires in Yosemite
National Park, USA. Fire Ecol 3: 34-52.
VESTRA 2003. HFQLG Vegetation Mapping Project Final Report. VESTRA Resources, Inc.,
Redding, CA, USA. p. 19.
22
Table 1. Characteristics of the five fires sampled and field sampling summaries in the northern
Sierra Nevada, California, USA.
Fire name year
Lookout - 1999
Pidgeon - 1999
Total
area
(ha)
StandTotal number of
Average stand- Number
replacing stand-replacing
replacing patch of field
1
proportion patches / no. sampled size (ha)
plots
12.9
1058
0.12
9/4
40
13.7
1907
0.09
11 / 7
31
Plot
elevation
range (m)
1507 - 1583
1591 - 1791
Bucks - 1999
13,855
0.10
96 / 4
61.0
60
1487 - 1695
Storrie - 2000
20,067
0.33
104 / 1
223.0
76
1254 - 1558
2591
0.28
27 / 10
26.1
70
1017 - 1912
Rich - 2008
1
Minimum patch size was 2 ha
23
Table 2. Weather parameters for fire simulations using RANDIG. Parameters were obtained
from the Cashman Remote Automated Weather Station. Wind speeds and directions represent
the dominant values from all 90th percentile and above observations. Fuel moistures represent the
97th percentile values for the predominant fire season in the area (June 1 – September 30).
Weather parameter
Winds
Fuel moisture (%)
1h
10 h
100 h
Live herbaceous
Live woody
Speed
(km h-1)
31
32
31
Value
Direction
(az.)
225
45
180
Relative
frequency
0.77
0.15
0.08
2
2
5
35
60
24
Figure 1. Fire perimeters, stand-replacing patches, and field plot locations for the five studied
fires in the northern Sierra Nevada, California, USA
25
Figure 2. Sampling design for all field plots.
26
Figure 3. Frequency distributions for sampled tree regeneration, by species group, for individual
plots, and averaged for stand-replacing patches. The density value on the horizontal axis
represents the upper bound for a given bin.
27
Figure 4. Frequency distributions for sampled shrub cover for individual plots, and averaged for
stand-replacing patches. The cover value on the horizontal axis represents the upper bound for a
given bin.
28
Figure 5. Regression tree output explaining the influence of the identified variables on plot-level
tree regeneration by species group. The length of the line from each split indicates the relative
proportion of total sum of squares explained by that split. Box and whisker plots displaying
regeneration density per hectare (log scale) and number of patches (n) are reported for the
resulting groups at each terminal node.
29
Figure 6. Core and buffered study area boundaries for the Meadow Valley area, Plumas National
Forest, California, USA. The actual treatments were implemented between 2003 and 2008, and
the theoretical treatments are based on output from the Treatment Optimization Model (Finney
2006). Weather data for fire modeling were obtained from both the Quincy and Cashman
Remote Automated Weather Station.
30
Figure 7. Conditional burn probabilities across the Meadow Valley landscape for which
simulated flame lengths are greater than 2 m. Burn probabilities are based on 10,000 randomly
placed ignitions simulated using RANDIG. Treated areas under the two treatment scenarios are
displayed.
31
Figure 8. . Mean conditional burn probabilities over time across the Meadow Valley landscape
for which simulated flame lengths are greater than 2 m. Probabilities are based on 10,000
randomly placed ignitions simulated using RANDIG.
32
Chapter 2:
Terrestrial Bird Module
Ryan D. Burnett, Paul Taillie, Nathaniel Seavy
Executive Summary
In 2009 the avian module of the Plumas-Lassen Area Study (PLAS) expanded to address
important questions related to post-fire habitat and its management. The primary objective of this
study is to assess the influence of post-fire conditions on spatial and temporal variation in bird
abundance and to use this information to guide forest management actions of post-fire
environments. In 2011 we continued sampling three areas that have burned in recent years within
the boundaries of the original PLAS study: the Storrie, Moonlight, and Cub fires, as well as 40
transects in the adjacent PLAS green forest treatment units.
We compared the abundance of 36 species that breed in the study area - representing a
range of habitat types and conditions - between the three burned areas and green forest study
sites. Sixteen species were significantly more abundant in burned areas and 14 were significantly
more abundant in unburned habitat.
Shannon index of diversity, species richness, and total bird abundance were significantly
higher in PLAS green forest in 2011 than any of the three burned areas. Among the three burned
areas, Storrie had the highest indices for each of these three metrics. All three metrics were
significantly lower in both the Moonlight and Storrie private lands than in the adjoining Forest
Service lands.
Burn severity was a significant predictor of avian diversity and total bird abundance
across burned areas with values highest at moderate burn severity, though these point level
measures of burn severity explained a relatively small portion of the variance in avian metrics.
We found and confirmed 122 active nests of 11 cavity nesting species in 2011 for a total
of 366 nests found from 2009 – 2011. We found 14 Black-backed Woodpecker nests in 2011 for
a total of 34 from 2009 – 2011. Black-backed and Hairy Woodpecker both showed strong
selection for high snag densities (>8 per 0.1 acre plot) surrounding nests.
Here we present 24 management recommendations, which are a culmination of our
results, scientific literature, and expert opinion. Some of these are hypotheses that should be
tested to ensure the best possible management practices are being employed to sustain avian
communities in the Sierra Nevada.
•
•
•
•
Post Fire Habitat Management Recommendations
General
Whenever possible restrict activities that depredate breeding bird nests and young to the nonbreeding season (August - April)
Consider post-fire habitat as an important component of the Sierra Nevada ecosystem because it
maintains biological diversity
Consider the area of a fire that burned in high severity, as opposed to the area of the entire fire,
when determining what percentage of the fire area to salvage log
Consider the landscape context (watershed, forest, ecosystem) and availability of different
habitat types when planning post-fire management actions
33
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Consider that snags in post-fire habitat are still being used by a diverse and abundant avian
community well beyond the 5 to 10 year horizon often suggested
Snags
Manage a substantial portion of post-fire areas for large patches (10 - 50 acres) burned with high
severity as wildlife habitat
Retain high severity burned habitat in locations with higher densities of larger diameter trees
Retain high severity patches in areas where pre-fire snags are abundant as these are the trees
most readily used in the first five years after a fire
Retain snags in salvaged areas far greater than green forest standards and retain some in dense
clumps
Snag retention immediately following a fire should aim to achieve a range of snag conditions
from heavily decayed to recently dead in order to ensure a longer lasting source of snags for
nesting birds
When reducing snags in areas more than five years post fire (e.g. Storrie fire) snag retention
should favor large pine and Douglas-fir, but decayed snags of all species with broken tops should
be retained in recently burned areas
Retain snags (especially large pine trees that decay slowly) in areas being replanted as they can
provide the only source of snags in those forest patches for decades to come
Consider retaining smaller snags in heavily salvaged areas to increase snag densities as a full size
range of snags are used by a number of species for foraging and nesting from as little as 6 inches
diameter, though most cavity nests were in snags over 20 inches
Early Successional Habitat
Manage post-fire areas for diverse and abundant understory plant community including shrubs,
grasses, and forbs. Understory plant communities provide a unique and important resource for a
number of species in a conifer-dominated ecosystem
Most shrub patches should be at least 10 acres, and shrub cover should average over 50% across
the area in order to support area sensitive species such as Fox Sparrow
Retain natural oak regeneration with multiple stems (avoid thinning clumps), as these dense
clumps create valuable understory bird habitat in post-fire areas 10 – 15 years after the fire
When treating shrub habitats, ensure some dense patches are retained. In highly decadent shrub
habitat consider burning or masticating half the area (in patches) in one year and burning the rest
in the following years once fuel loads have been reduced.
Maximize the use of prescribed fire to create and maintain chaparral habitat and consider a
natural fire regime interval of 20 years as the targeted re-entry rotation for creating disturbance
in these habitat types
Shaping Future Forest
Limit replanting of dense stands of conifers in areas with significant oak regeneration, and when
replanting these areas use conifer plantings in clumps to enhance the future habitat mosaic of a
healthy mixed conifer-hardwood or pine-hardwood stand
Consider managing smaller burned areas (<5000 acres) and substantial portions of larger fires
exclusively for post-fire resources, especially when there have been no other recent fires in the
adjoining landscape.
34
•
•
•
•
•
•
Retain patches of high burn severity adjacent to intact green forest patches, as the juxtaposition
of unlike habitats is positively correlated with a number of avian species, including those
declining such as Olive-sided Flycatcher, Western Wood-Pewee, and Chipping Sparrow
Incorporate fine scale heterogeneity in replanting by clumping trees with unplanted areas
interspersed to create fine scale mosaics and invigorate understory plant communities and natural
recruitment of shade intolerant tree species
Plant a diversity of tree species where appropriate, as mixed conifer stands generally support
greater avian diversity than single species-dominated stands in the Sierra Nevada
Consider staggering plantings across decades and leaving areas to naturally regenerate in order to
promote uneven-aged habitat mosaics at the landscape scale
Consider fuels treatments to ensure the fire resiliency of remnant stands of green forest within
the fire perimeter, as these areas increase avian diversity within the fire and the edges between
unlike habitats support a number of species (e.g. Olive-sided Flycatcher)
Avoid planting conifer species in or adjacent (depends on the size of riparian corridor) to riparian
areas to avoid future shading of riparian deciduous vegetation and desiccation. Consider
replanting appropriate riparian tree species (cottonwood, willow, alder, aspen) where appropriate
Introduction
In the Sierra Nevada, there is a pressing need to understand the nexus of silvicultural
practices, wildfire, and fuels treatments in order to maintain forest ecosystems that are
ecologically diverse and resilient. In the context of a century of fire suppression, at the core of
the debate over how to manage Sierra forests is how to most appropriately manage areas where
natural disturbance regimes have been disrupted. Land managers need more information about
the suitability of habitat created through fire suppression, fuel treatments (DFPZ, groups,
mastication), and wildfire and post-wildfire management to ensure the goals of maintaining
biological diversity are achieved.
The challenge of integrating wildfire and forest management into wildlife conservation is not
unique to the Sierra Nevada. Because large, infrequent disturbances are responsible for longlasting changes in forest structure and composition (Foster et al. 1998), they are recognized as a
critical element of bird community dynamics (Brawn et al. 2001). In many regions of western
North America, fires burn with considerable spatial and temporal variability (Agee 1993),
creating complex mosaics of vegetation patches. In these systems, changes in bird abundance are
often linked to post-fire vegetation characteristics and landscape composition (Saab et al. 2002,
Huff et al. 2005, Smucker et al. 2005).
In addition to fire suppression, there are a number of management activities that influence
post-fire vegetation characteristics and landscape composition in working forests. These
activities include salvage-logging, the mechanical mastication and herbicidal treatments to
reduce broadleaf shrubs, and planting of conifer species that are favored by forestry. As a result,
management activities may have profound influences on post-fire conditions locally and across
landscapes.
Beginning in 2009 the avian module of the Plumas-Lassen Administrative Study (PLAS)
expanded to address important questions related to post-fire habitat and its management. The
primary objective of this study is to assess the influence of post-fire conditions on spatial and
temporal variation in bird abundance and to use this information to guide forest management that
can maintain avian diversity across multiple spatial scales. We began sampling three areas
35
affected by fire within the boundaries of the original PLAS study: the Storrie fire that burned in
the fall of 2000, the Moonlight fire that burned in the Fall of 2007, and the Cub fire that burned
in the summer of 2008. Each of these fires burned at similar elevations and through primarily
mixed conifer and true fir vegetation communities, but with varying severity patterns. This report
provides results from the first three years of avian monitoring and uses ongoing monitoring of
unburned, actively managed “green” forest in the study area to provide context.
Methods
Study Location
The Plumas-Lassen Area Study avian module encompasses portions of the Mount Hough
Ranger District of Plumas National Forest and the Almanor Ranger District of Lassen National
Forest in the Sierra Nevada Mountains of northeastern California (Figure 1). In 2009 we added
three separate burned areas to our study within this same area. The elevations of sites surveyed
ranged from 1126 to 1998 m with a mean of 1658 m in the Cub fire, 1199 to 2190 m with a mean
of 1779 m in the Moonlight fire, 1107 to 2011 m with a mean of 1528 m in the Storrie fire, and
1094 to 1902 m with a mean of 1483 m at the existing PLAS green forest sites.
Site Selection
Random starting points for each burned transect were generated in ArcGIS 9.2 within the
boundaries of each fire (ESRI 2004). The original sampling area was limited to Forest Service
land and sites with a slope of less than 40 degrees to allow access and safe navigation on foot in
a timely manner. We maintained a minimum distance between transect starting points of 1500 m
to ensure transects would not overlap and to maximize a spatial balance within the sampling
frame of each fire. Four more points were added to the starting point on a random compass
bearing at 250 m spacing resulting in a 1 km long, five-point transect. We minimized the number
of point counts to allow ample time to conduct cavity nest searches during the prime morning
hours when bird activity is greatest.
Even with stratification to eliminate areas with steep slopes, we had to drop
approximately 10% of the original transects selected following field reconnaissance due to issues
with access, safe navigation, or noise from nearby streams. We replaced the majority of these
transects with new locations until nearly all surveyable areas in each fire were covered.
However, due to steep topography and large roadless areas in the Storrie fire – and to a lesser
degree the Cub fire – our sampling is not evenly distributed across these fires as it is in the
Moonlight.
In 2010 we added an additional ten transects on private land managed by W. M. Beaty &
Associates, Inc; six in the Moonlight Fire, and four in the Storrie fire. These transects were
selected using the same protocol as described above, but in this case, the sample area was
defined as the private land within the fire boundary, shown in white on the map (Figure 1). We
then selected the maximum number of transects that the sample area size would allow.
A total of 60 transects (300 stations) were surveyed across the three fires in 2011. A total
of 32 transects were surveyed in the Moonlight fire (including 6 on private land), 13 in the Cub
Fire, and 15 in the Storrie fire (including 4 on private land). Snow prevented access to two
Storrie transects (ST04 and ST15) before July, thus they were not surveyed in 2010 or 2011. Site
selection for PLAS green forest study sites followed a similar random selection protocol except
each transect contained 12 points instead of five, and approximately 25% of transects were
systematically established in areas where treatments were planned (many now implemented).
36
The PLAS site selection protocol for the unburned “green forest” sample is described in detail in
the original PLAS study plan and previous annual reports (Stine et al. 2005, Burnett et al. 2009).
37
Figure 1. The location of PRBO point count transects in the Plumas-Lassen study area in 2010 & 2011.
38
Figure 2. Location of PRBO point count stations overlaid on composite burn index fire severity maps for each of
three fires in the study area. Red = high severity, Orange = moderate, Lime = low severity, and green is unburned.
Storrie
39
Cub
40
Moonlight
Bird community surveys
The avian community was sampled using a five minute, exact-distance, point count census
(Reynolds et al. 1980, Ralph et al. 2005). In this method points are clustered in transects, but data
are only collected at the individual point. All birds detected at each point during the five-minute
survey were recorded according to their initial distance from the observer. The method of initial
detection (song, visual, or call) for each individual was also recorded. All observers underwent an
intensive, three week training period focused on bird identification and distance estimation prior to
conducting surveys. Laser rangefinders were used to assist in distance estimation at every survey
point. Counts began around local sunrise, were completed within four hours, and did not occur in
inclement weather. Aside from the two transects that were not accessible for the entire season
(ST04 and ST15), all but two transects were visited twice during the peak of the breeding season
from mid-May through the first week of July. Cub 7 and Cub 8 were only visited once in 2011 due
to administrative forest closures in those areas.
Cavity nest surveys
In addition to point count censuses, at each fire transect a 20 ha area (200 x 1000 m
rectangle) surrounding the point count stations was surveyed for nests of cavity-nesting birds
following the protocol outlined in “A field protocol to monitor cavity-nesting birds” (Dudley and
Saab 2003). In order to focus our attention on species of interest we ignored some of the more
common cavity-nesters. Our focal species included both species of bluebird, all woodpeckers, and
all cavity-nesting raptors.
After the point count census was complete, the nest survey was conducted for between two
and four hours depending on the habitat, terrain and time spent waiting to confirm a cavity’s status.
41
All nest surveys were completed by noon. The primary search method for finding nests was bird
behavior, though once suspicious birds were located observers often conducted a systematic search
of snags in the vicinity. Once a potential nest was found, it was observed from a distance for
approximately 20 minutes to confirm the cavity was an active nest. If that cavity was confirmed
active, a variety of characteristics of both the nest tree and the cavity were recorded. These
characteristics included diameter at breast height (DBH), tree height, tree species, cavity height,
tree decay class, and the orientation of the cavity opening. For tree decay, we used a qualitative
scale of decay ranging from one to eight, with one being a live, intact tree and eight being a
severely decayed stump (Appendix C). If the observer was unable to confirm the cavity was
active, its location was recorded to aid nest searchers during the second visit. Only confirmed
active nests were used in analysis presented herein.
Figure 3. PRBO Northern Sierra post-fire habitat survey plots.
Vegetation surveys
Vegetation data was collected at all burned points in 2009, 2010, and 2011. We measured
vegetation characteristics within a 50 m radius plot centered at each point count station following
the relevé protocol outlined in Stine et al. (2005). On these plots we measured shrub cover, live
tree cover, and herbaceous cover as well as the relative cover of each species in the shrub and tree
layers through ocular estimation. We also collected basal area of live trees and snags using a 10factor basal area key. To estimate the density of snags across the plot, we recorded data (e.g. DBH,
species, height, decay) on every snag within 11.3 m of the center of the point count location. In
addition to the point count stations, we collected the same snag data at all active nests, as well as at
5 random locations distributed throughout the 20 ha nest plot. To select these random points,
coordinates were first generated in ArcGIS 9.2 to serve as a guide (ESRI 2004). Once in the field,
the observer navigated to within 10 m of the random point and then chose the closest tree over
12 cm DBH as the center of the random snag plot. The center trees of these random snag plots
42
were used as a sample of “random nest trees,” and all data collected for active nests were also
collected for these random “nests.”
Analysis
We compared an index of the abundance of 36 species (detections within 50 m of observer)
on Forest Service land across all three fires combined and the PLAS green forest. The index was
derived by taking the summed detections per year (2 visits to each site) within 50 m of observers.
We then used negative binomial regression with total count of each species as the independent
variable and burn status as the dependent variable. We limited our analysis to species with a mean
per point abundance for all fires combined or all green forests sites combined of at least 0.10. We
used this cutoff as previous analysis with these data suggested it was a good estimate of the level
for which we had enough power to detect significant effects. We also included Black-backed
Woodpecker and Mountain Quail – species that did not meet this abundance threshold – as they
are Management Indicator Species for National Forest lands in the Sierra Nevada. This resulted in
a list of 36 species which represented 95% of all detections within 50 meters of observers in our
dataset. We present mean detections within 50 m of observers per point count station per visit to be
comparable to previous reports using these data, but analysis was conducted on total detections to
meet the assumptions of negative binomial regression (Cameron and Trivedi 1998).
In order to quantify the overall songbird community in the study areas, we used three
different metrics: the Shannon Index of species diversity, species richness, and total bird
abundance. The Shannon index used a transformation of Shannon’s diversity index (or H', Krebs
1989) denoted N 1 (MacArthur 1965). The transformation expresses the data in terms of number of
species and thus is more easily interpreted. Expressed mathematically:
i =S
N = e H ' and H ' =
( pi )(ln pi )(-1)
1
∑
i =1
where S = total species richness and p i is the proportion of the total numbers of individuals for
each species (Nur et al. 1999). High Shannon index scores indicate both high species richness and
more equal distribution of individuals among species. Species richness is defined simply as the
number of species detected within 50 m of each point summed across the two visits, and total bird
abundance is the sum of all species detected per visit within 50 m. All species that do not breed or
naturally occur in the study area and those that are not adequately sampled using the point count
method – including waterfowl, shorebirds, waders, and raptors – were excluded from each
calculation. These metrics were investigated for each fire and the Plumas Lassen Administrative
Study green forest study sites (Figure 4).
We investigated the effect of burn severity on avian diversity and total bird abundance
across each of the three fires and all fires combined. To classify burn severity we used a relative
differenced normalized burn ration (RdNBR) ground-truthed and converted to Composite Burn
Index (CBI) units using regression (Miller and Thode 2007). This index represents the magnitude
of effects caused by the fire and incorporates various strata including changes to soil, amount of
vegetation and fuels consumed, re-sprouting from burned plants, and blackening or scorching of
trees (Key and Benson 2005). The values of CBI were derived from the Relative Differenced
Normalized Burn Ratio (RdNBR) which was in turn derived from Landsat Thematic Mapper
imagery, a method discussed by Miller and Thode (2007). The values range from zero,
representing no sign of burning, to three, representing the highest severity. The Landsat data used
to derive CBI were collected on 23 July 2008 for the Cub, 7 July 2008 for the Moonlight, and 21
August 2001 for the Storrie fire and were provided by the Forest Service.
43
We excluded all sites that had been treated (e.g. salvaged or masticated) and all private
lands. We used linear regression with untransformed bird data to evaluate the effect of CBI on
these metrics. We standardized values of CBI in order to evaluate non-linear relationships of the
effect of fire severity on avian community metrics expressed as:
where x = each point level estimate of CBI, µ = the sample mean, and σ = standard deviation of the
sample.
For each metric we ran the models again with a squared standardized CBI term and
compared the goodness of fit between the linear and 2nd order models using a likelihood ratio test.
For each best fit model we present F-statistics and p-values and for goodness of fit tests we present
χ2 likelihood ratio statistics and associated p-values in Appendix A.
We compared an index of abundance of the 30 species that showed a preference for burned
or unburned forest between salvaged private land and un-salvaged National Forest land in the
Moonlight and Storrie fires. We compared the summed total detections of all species in each group
averaged per point count station across 2010 and 2011. We used detections within 50 m of
observers to reduce any biases in detectability. We evaluated the differences in abundance using a
two-tailed t-test with the assumption of unequal variance and assumed statistical significance at the
p < 0.05 level. All point count data collected for this study are available on the Sierra Nevada
Avian Monitoring Information Network: http://data.prbo.org/apps/snamin/index.php?page=firehome-page .
We evaluated selection for snag densities surrounding nest trees for the six cavity nesting
species for which we had adequate sample sizes (Black-backed, Hairy, and White-headed
woodpeckers, Mountain Bluebird, Northern Flicker, and Red-breasted Sapsucker) in the Cub and
Moonlight fires. We excluded the Storrie fire due to lack of sample size for both Black-backed and
Hairy Woodpeckers. We considered all randomly-selected trees to be available for nesting, such
that we were comparing a sample of available nest trees to a sample of the used nest trees. This
corresponds to the Sampling Protocol A and Design I described by Manly et al. (2002). We pooled
data across the three years of the study. Because we searched the same plots in all years, it was
likely that the same pairs were recorded in multiple years. This corresponds to Design 1 and is
appropriate for making inferences about the population in our study area (Manly et al. 2002).
We evaluated snag density for each burn separately as well as combined. For each of the
three snag density categories, we calculated the proportion of used and available trees. We treated
both the used and available trees as a sample of the larger population and calculated the standard
error of the proportion in each category following Manly et al. (2002). After preliminary analyses
revealed relatively minor differences between the two burns, we limited our subsequent analyses
of selection ratios to the pooled data set.
We calculated selection ratios by dividing the proportion of a used resource by the
proportion of the corresponding available resource. If nest trees were used in proportion to their
availability they would have a selection ratio of 1.0. A selection ratio > 1 implied preference, while
a value < 1 implied avoidance. We first used a Pearson’s chi-square goodness-of-fit test to evaluate
if the pattern of nest tree use differed from the pattern of available trees (Manly et al. 2002). If we
found evidence of preference (P < 0.05), we further interpreted the selection ratios for each
category using simultaneous 95% Bonferroni confidence interval calculated over all categories. If
these confidence intervals did not include 1, we rejected the null hypothesis of use proportional to
44
availability (Manly et al. 2002). Nest site snag density selection analysis was conducted in the
program R (R Version 2.10.1, http://cran.r-project.org/, accessed 27 November 2011), using the
package adehabitat (Calenge 2006).
Results
Avian Community Composition
Comparing the abundance of 36 species that breed in the study area and represent a range
of habitat types and conditions, 16 were significantly more abundant in burned areas, 14 in
unburned areas, and 6 showed no statistical difference (Table 1). Of the six showing no difference,
each reached their greatest abundance in one of the three burns.
The patterns in species distribution observed in 2009 and 2010 continued in 2011, with
species associated with low severity fire and unburned green forest being among the most
abundant in the Cub Fire while in the Moonlight and Storrie fires species more strongly associated
with early successional habitats and disturbance (e.g. woodpeckers, shrub nesters) were among the
most abundant (Table 1). One exception was Black-backed Woodpecker, which was found in
relatively small patches of high snag density throughout the Cub fire.
Shannon index of avian diversity, species richness, and total bird abundance continued to
show similar patterns in 2011. Each index was highest in unburned forest and lowest on private
land (Figure 4). All three indices increased across each of the three fires and the PLAS green forest
in 2011 for the second consecutive year, with Moonlight showing the greatest increases; in 2011
indices there were the highest for any of the three fire areas monitored.
45
Table 1. An index of the abundance (mean detections per point per visit < 50m) of 36 species on National Forest
land in three burned areas and the adjacent unburned Plumas-Lassen Area Study (PLAS) averaged from 2009 2011.Species are listed based on whether they were significantly (p<0.05) more abundant within all burned areas
combined or the PLAS green forest study area. SE = Standard Error.
More Abundant Inside Burn
Mountain Quail
Hairy Woodpecker
White-headed Woodpecker
Black-backed Woodpecker
Olive-sided Flycatcher
Western Wood-Pewee
Mountain Bluebird
Brown Creeper
House Wren
Lazuli Bunting
Spotted Towhee
Green-tailed Towhee
Chipping Sparrow
Fox Sparrow
Dark-eyed Junco
Cassin's Finch
More Abundant Outside Burn
Hammond's Flycatcher
Dusky Flycatcher
Cassin's Vireo
Mountain Chickadee
Red-breasted Nuthatch
Golden-crowned Kinglet
Hermit Thrush
Nashville Warbler
Yellow-rumped Warbler
Hermit Warbler
Western Tanager
Black-headed Grosbeak
Evening Grosbeak
Pine Siskin
No Statistical Difference
Calliope Hummingbird
Warbling Vireo
Stellar’s Jay
American Robin
Yellow Warbler
MacGillivray's Warbler
Cub
0.01
0.12
0.09
0.02
0.05
0.06
0.01
0.23
0.01
0.02
0.04
0.05
0.02
0.12
0.37
0.05
SE
0.01
0.02
0.02
0.01
0.01
0.02
0.01
0.03
0.01
0.01
0.01
0.01
0.01
0.03
0.04
0.01
Moonlight
0.01
0.11
0.05
0.03
0.04
0.09
0.09
0.15
0.06
0.32
0.04
0.06
0.17
0.35
0.72
0.11
SE
<0.01
0.02
0.01
0.01
0.01
0.01
0.01
0.02
0.01
0.03
0.01
0.01
0.02
0.03
0.04
0.01
Storrie
0.03
0.03
0.03
0.00
0.07
0.13
0.01
0.13
0.25
0.37
0.33
0.09
0.15
0.55
0.40
0.06
SE
0.02
0.01
0.01
0.00
0.02
0.03
0.01
0.03
0.04
0.05
0.04
0.02
0.02
0.06
0.04
0.02
PLAS
0.004
0.03
0.02
0.002
0.03
0.03
0.001
0.13
0.04
0.03
0.07
0.01
0.03
0.22
0.45
0.03
SE
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
0.01
0.01
<0.01
0.01
0.02
<0.01
0.01
0.01
<0.01
0.18
0.16
0.05
0.40
0.42
0.12
0.01
0.08
0.45
0.30
0.27
0.07
0.06
0.04
0.04
0.03
0.01
0.04
0.04
0.02
<0.01
0.02
0.04
0.05
0.03
0.02
0.02
0.02
0.14
0.36
0.03
0.26
0.13
0.07
0.02
0.05
0.33
0.15
0.33
0.03
0.01
0.09
0.02
0.03
0.01
0.02
0.02
0.01
0.01
0.01
0.03
0.02
0.02
0.01
<0.01
0.02
0.03
0.25
0.04
0.23
0.17
0.06
<0.01
0.38
0.20
0.13
0.11
0.10
0.08
0.03
0.01
0.03
0.01
0.03
0.03
0.01
<0.01
0.04
0.03
0.02
0.02
0.02
0.02
0.01
0.26
0.36
0.21
0.42
0.43
0.33
0.06
0.40
0.48
0.89
0.32
0.11
0.14
0.11
0.01
0.02
0.01
0.01
0.02
0.01
<0.01
0.02
0.01
0.02
0.01
0.01
0.01
0.01
0.01
0.09
0.10
0.05
0.03
0.06
0.01
0.02
0.02
0.01
0.01
0.02
0.05
0.10
0.06
0.10
0.03
0.17
0.01
0.02
0.02
0.01
0.01
0.02
0.06
0.04
0.07
0.07
0.15
0.23
0.01
0.01
0.02
0.02
0.03
0.03
0.04
0.10
0.08
0.07
0.05
0.15
<0.01
0.01
<0.01
<0.01
<0.01
0.01
46
Figure 4. An index of avian richness, ecological diversity (Shannon index), and total bird abundance from
detections within 50m of point count stations on National Forest and private land in the Storrie, Moonlight, and
Cub fires and Plumas unburned forest (PLAS)
Species Richness
Species/Point
10.00
8.35
8.00
6.00
5.83
6.65
6.47
4.83
4.60
4.00
2.00
0.00
Cub
Moonlight
Moonlight
Private
Storrie
Storrie
Private
PLAS
Species/Point (weighted)
Shannon Diversity
8.00
7.00
6.00
5.00
4.00
3.00
2.00
1.00
0.00
5.69
6.39
4.31
4.16
Cub
Moonlight
7.42
6.38
Moonlight
Private
Storrie
Storrie
Private
PLAS
Detections/ Point visit
Total Bird Abundance
8.00
7.00
6.00
5.00
4.00
3.00
2.00
1.00
0.00
6.94
4.55
Cub
5.43
5.33
4.09
3.58
Moonlight
Moonlight
Private
Storrie
Storrie
Private
PLAS
47
Effect of Burn Severity on Avian Community Metrics
Burn severity had significant effects on avian community indices but explained a
relatively small portion of the variance observed (Figure 5). Burn severity had a significant effect
on avian diversity for the Cub, Moonlight, and all fires combined. The fit was significantly
improved by including a squared term for burn severity for each, with avian diversity highest at
sites that burned at moderate severity. Burn severity did not have a significant effect on avian
diversity in the Storrie fire, though there was some evidence (p=0.12) of a similar quadratic
effect. Burn severity explained relatively a small portion of the variance in species diversity for
all fires combined and Moonlight (~6%). For the Cub fire model performance was better, with
21% of the variance explained by severity. For all fires combined the effect of severity on the
abundance of all species combined was significant, with a squared term improving fit and
abundance greatest at moderate severity. However, the Storrie and Moonlight fires showed
significant linear but opposite effects of severity, with Storrie total abundance increasing with
increased severity and Moonlight decreasing. The effect of severity was not a significant
predictor of total bird abundance in the Cub fire. The variance in total bird abundance explained
by burn severity was relatively small for all fires combined and Moonlight (~6%) but better for
the Storrie fire (14%). Though there was some evidence of an effect of severity on these metrics
for private land, with the relatively small sample size (n=30 & n=20) and lack of variation in fire
severity on these lands we considered fitting a regression inappropriate. When data from private
land on both the Storrie and Moonlight fires were combined there was no effect of severity for
avian diversity (p=0.44) or abundance (p=0.80) on private land (P=0.80).
Salvaged vs. Un-salvaged
The extensively salvaged logged private land within both the Moonlight and Storrie fires
had significantly lower avian diversity and abundance in 2010 and 2011 than adjacent Forest
Service land (Figure 4). In the Moonlight Fire, private land averaged 2.05 fewer species per point
and 1.75 fewer total birds than Forest Service land. The pattern was similar for the Storrie fire,
where private land supported 1.36 fewer species and 1.75 total birds per point compared to
Forest Service land. Extrapolating these figures out to the acreage of private land in each fire
would result in approximately 17,000 fewer total individual birds in the roughly 19,000 acres of
private land in the Moonlight fire and 800 fewer individuals in the roughly 4,000 acres of private
land in the Storrie fire compared to similar acreage of Forest Service land in these fires.
48
Figure 5. The effect of burn severity (Composite Burn Index) on avian community indices per point count station
at three fires in the Northern Sierra from 2010 & 2011 with best fit line. All graphed relationships had a
significant effect of fire severity (p<0.05). See Appendix A for full regression models.
49
Ten of the 50 point count stations on private land randomly fell within or immediately
adjacent to (<30 m) riparian areas that were not subjected to the same mechanical salvage or
herbicide treatments as the surrounding uplands. Avian species richness and total bird
abundance were significantly higher within these riparian areas across both fires (Figure 6). The
values observed in riparian areas were similar to those on un-salvaged National Forest land. The
elevated avian community metrics in these areas was attributable to open-cup shrub-nesting
species, which were more than three times more abundant in riparian areas than they were in unsalvaged upland private land. Cavity nesters were not significantly more abundant in riparian
areas, though snag densities were higher and we did observe a number of Lewis’ Woodpeckers
using these areas in the Storrie fire.
An index of the total abundance of both burn and green forest associated species was
significantly greater on un-salvaged Forest Service land than salvaged private land (Figure 7).
The difference in abundance was far greater for green forest-associated species than it was for
burned forest associates. However, the index of total woodpecker abundance was more than 2.5
times greater on un-salvaged Forest Service land (0.48 per point) than on private land (0.18).
These results were very consistent across the two burns investigated.
Figure 6. Mean per point Shannon index of bird diversity and total bird abundance (<50m per point per visit) at
salvaged and un-salvaged private land and adjacent un-salvaged naturally regenerating National Forest land in
the Storrie and Moonlight fires in 2010.
Salvaged vs. Unsalvaged
8
7
Private Salvaged
Private Unsalvaged
Unsalvaged - USFS
6
5
4
3
2
1
0
Species Diversity
Total Bird Abundance
50
Figure 7. An index of the abundance (mean detections <50m per point per visit) of two species groups on
salvaged (private) and un-salvaged (National Forest) land in the Moonlight and Storrie fires in 2010 and 2011
with 95% confidence intervals. The first group includes all species who were significantly more abundant in
unburned areas in our study area and the second group is comprised of those species signifiantly more abundnt
in burned areas (following results in Table 1).
Un-Salvaged
Detections Per Point Count Visit
Salvaged vs. Un-Salvaged
Salvaged
3.5
3
2.5
2
1.5
1
0.5
0
Green Forest
Species
Burned Forest
Species
Moonlight
Green Forest
Species
Burned Forest
Species
Storrie
Cavity Nests
A total of 122 active cavity nests of 11 species were located in 2011 across the Cub,
Moonlight, and Storrie fires. Thirty-one nests were found in the Cub fire, 71 in the Moonlight
Fire, and 20 in the Storrie fire. In the Cub fire we found more nests for Hairy Woodpecker than
any other species, in the Moonlight fire Mountain Bluebird, and in the Storrie fire Red-breasted
Sapsucker (Table 2). For the third consecutive year we did not find Black-backed Woodpecker
nests in the Storrie fire, and for the first year we did not find any Hairy Woodpecker nests there
either.
There was evidence of selection for areas with the highest snag densities and avoidance
of areas with low snag densities surrounding nests for four of the six cavity nesting species we
had sufficient sample sizes to investigate in the Cub and Moonlight fires (Figure 8). Whiteheaded Woodpecker and Red-breasted Sapsucker showed no strong selection, though they
appear to be avoiding nesting in areas with the lowest snag densities. Mean snag density in the
0.1 acre surrounding nests was 13.3 (7.6 SD) for Black-backed Woodpecker, 9.8 (5.8 SD) for
Hairy Woodpecker, 6.8 (5.7) for White-headed Woodpecker, and 5.0 (5.4 SD) at randomly
selected locations.
51
Table 2. Number of cavity nests confirmed by species and fire in the Plumas-Lassen Study 2011 with totals from
2009 - 2011.
American Kestrel
Black-backed Woodpecker
Hairy Woodpecker
White-headed Woodpecker
Lewis' Woodpecker
Northern Flicker
Pileated Woodpecker
Red-breasted Sapsucker
Williamson's Sapsucker
Mountain Bluebird
Western Bluebird
Total
Cub
Moonlight
Storrie
Grand
2011 Total 2011 Total 2011 Total Total
0
0
1
2
0
0
2
6
15
8
19
0
0
34
11
26
7
41
0
5
72
3
13
6
26
6
17
56
0
0
2
7
0
7
14
5
13
8
27
4
18
58
3
4
0
1
0
0
5
2
2
10
25
7
11
38
0
0
1
3
0
0
3
1
4
24
61
1
4
69
0
0
4
8
2
7
15
31
77
71
220
20
69
366
The nest trees selected across the three fires were quite variable in both size and species,
though the vast majority of nests were found in trees greater than 50 cm (20 in) DBH (Figure 9).
There was more evidence for preference for tree decay, which is shown by both decay class and
top condition of the nest tree (Figure 10). The vast majority of the nests were in dead trees
(decay class 3 or greater) despite roughly a third of the “available” trees being alive. Hairy
Woodpecker, Black-backed Woodpecker, and Red-breasted Sapsucker chose mostly trees of
decay class 3 and 4, while White-headed Woodpecker, Northern Flicker, and Lewis’
Woodpecker chose trees of higher decay classes. Most Woodpeckers showed preference for
broken top trees, with the one exception being Black-backed Woodpecker, which chose trees
with top condition similar to availability. Black-backed Woodpeckers appear to be using the least
decayed and most intact snags of any woodpecker species across these fires. All of these patterns
were consistent across the three years of monitoring.
52
Figure 8. Snag densities within an 11.3 m radius plot around nest trees for six species in two burns compared to available sites. The upper panel is the
proportion of all nest trees and randomly-selected trees with 95% confidence intervals. The p-values are from Pearson’s chi-squared goodness-of-fit tests for
the null hypothesis that the proportion of nest trees did not differ from the available. Bottom panel presents selection ratios, if nest trees were used in
proportion to their availability they would have a selection ratio of 1; a selection ratio > 1 implies preference, whereas a value < 1 implies avoidance.
53
Figure 9. Median (dot), 25-75% interquartile range (box), and extremes (bars) in nest tree diameter at breast
height (DBH) and nest tree height for six woodpecker species and randomly selected trees in the Storrie,
Moonlight, and Cub fires in 2009 - 2011.
Nests were found in eight tree species/species groups from 2009 to 2011 (Figure 11).
However, the majority of nests were found in five species/species groups: true fir (both red and
white), yellow pine (includes both Ponderosa and Jeffrey), and Douglas-fir. These were some of
the most common tree species available based on our sample of random trees. However, incense
cedar was also one of the more common available trees, though only one nest (White-headed
Woodpecker) was found in a cedar. Black-backed Woodpecker almost exclusively selected true
fir and pine species, while Hairy Woodpecker, White-headed Woodpecker, and Northern Flicker
were less particular, with the relative frequencies of use reflecting what was available. Over a
quarter of the Red-breasted Sapsucker nests were found in aspen despite this tree species having
an extremely small proportion of the available trees. Similarly, Lewis’ Woodpecker chose
Douglas-fir as a nest tree more than it was available, while none of the nine Lewis’ Woodpecker
nests were in true fir, the most prevalent tree species.
54
Figure 10. The decay class and top condition of the nests of 6 woodpecker species compared to the sample of
randomly selected trees across three fires in the Northern Sierra Nevada from 2009 - 2011.BA=Broken After Fire,
BB=Broken Before Fire, DT=Dead Top, F= forked, I= Intact.
55
Figure 11. Figure 11. Nest tree species for 6 woodpecker species compared to the sample of random trees across
three fires in the Northern Sierra Nevada from 2009 - 2011. ABISPP = true fir, CALDEC = incense cedar, PINCON =
lodgepole pine, PINLAM = sugar pine, PINMON = western white pine, PINPON = ponderosa pine, POPTRE =
quaking aspen, PSEMEN = Douglas-fir, QUEKEL = black oak, SEQGIG = giant sequoia.
Discussion
Avian Community Composition in Burned vs. Green Forest
The vast majority of species breeding in our study area showed a preference for either
burned or unburned forest. This illustrates the profound effect that fires have on habitat
composition and hence distribution of biological diversity in the Sierra Nevada. This dichotomy
in the avian community suggests the need for different management strategies in burned areas
compared to those developed for green forests in the Sierra Nevada. For example, several
woodpecker species in the study area are avoiding nesting in areas with snag densities less than
40 per acre. Thus, green forest snag retention guidelines of 4 per acre should be considered
inappropriate for sustaining habitat for cavity nesting species in post-fire areas (Hutto 2006).
Many of the species more abundant in post-fire areas are associated with habitat components
such as high snag densities or dense shrub layers, herbaceous vegetation, and regenerating
hardwoods (e.g., oak and aspen). Snag retention guidelines, broad leaf herbicides or mastication,
and ground disturbance may have significant effects on the avian community in post-fire
environments. While some snag associated species (e.g. Black-backed Woodpecker) decline five
56
or six years after a fire, those associated with understory plant communities take their place,
resulting in similar avian diversity three and eleven years after fire (e.g. Moonlight and Storrie).
An understanding of the differences in avian community composition and specifically the
relative abundance of species between unburned forest, mechanical fuels reductions, and postfire habitats can help guide the management of these areas. With high severity burns having
lower densities of late seral associated species, these areas might best be prioritized for
sustaining populations of early successional species. Likewise, later seral habitat areas are
probably not the ideal location for creating large quantities of disturbance-dependent habitat
elements. Ensuring areas valued for their late seral habitat characteristics are resilient to standreplacing fire seems a far better strategy for managing for species like the Spotted Owl than
eliminating early successional habitat in burned areas in order to speed the return of mature
forest. However, habitat mosaics are also an important part of ensuring biological diversity in
these forests; thus a science-driven adaptive management approach to strategic salvage and
reforestation in large fires that burned at high severity (e.g. Moonlight) may result in long term
benefits to avian diversity. A greater understanding of appropriate patch sizes and configuration
is needed to guide this process both in terms of the negative impacts of salvage and positive
impacts of creating uneven aged future stand heterogeneity.
Time Since Fire and Burn Severity
The effects of burn severity on the avian community varied across the three fires. Factors
such as time since fire, landscape patterns of severity, and pre-fire habitat conditions are likely
interacting with severity to influence the avian community at the 50 m scale we sampled birds
and the pixel-level scale we sampled fire severity (Saab et al. 2007, 2009). Hence, a relatively
small portion of the variation in avian metrics was explained by burn severity alone even though
models were highly significant. One pattern that appears fairly consistent across the three fires is
a non-linear (quadratic) effect of burn severity on avian species diversity, with highest values in
the moderate severity categories. Further investigation into severity patterns at the landscape
scale is necessary to understand if it is moderate severity specifically or if moderate severity is
potentially a good predictor of mixed severity in the larger surrounding landscape. Across each
of the three fires, the lowest avian diversity was recorded for the highest severity class, but for
the Cub and Storrie fires the unburned areas within the fire perimeter also had relatively low
avian diversity. These results suggest that managing for mixed severity wildfire with the majority
of the landscape burning at moderate severity (1 – 2 on the CBI scale) may maximize local and
landscape level avian diversity. In 2012 we plan to evaluate the local and landscape level
predictors of avian distribution across these three burned areas, including covariates of time since
fire, patch size and configuration, and pre-fire habitat conditions.
Managing for dense and diverse shrub habitats interspersed with areas of green forest
should maximize avian diversity in post-fire environments. Information about colonization rates
and how long after a fire shrub-dependent species persist at maximum levels can be used to
determine appropriate re-entry rotations for managing habitat following fire. Based on our results
and observations made of habitat conditions within the fires, there is a five year lag before dense
shrub habitats form that maximize densities of species such as Fox Sparrow, Dusky Flycatcher,
and MacGillivray’s Warbler. These species have shown substantial increases in abundance in the
Moonlight fire each year since 2009, but shrub-nesting species are still more abundant in the
eleven year post-burn Storrie fire. This suggests early successional shrub habitats in burned areas
provide high quality habitat for shrub-dependent species well beyond a decade after fire. A re-
57
entry rotation of 20 - 30 years for managing chaparral habitat may maximize abundance of these
species. This re-entry timeframe would mimic the historic fire return interval for montane
chaparral habitat in the Sierra Nevada (Barbour and Major 1988).
Results from the PLAS green forest study suggest that the use of prescribed fire has far
more positive effects on the avian community compared to the use of mechanical mastication in
shrub habitats in the region (Burnett et al. 2009). If mechanical mastication is used, especially in
high quality shrub bird habitat – as currently exists in the Storrie fire – retaining leave islands of
very dense shrubs will help provide nesting habitat and reduce negative impacts to shrubdependent species. However, best management practices for these species would be to avoid
disturbing this habitat for a number of years beyond the eleven since the Storrie fire burned.
Post-fire Cavity Nest Characteristics
The importance of post-fire habitats for cavity-nesting and bark-foraging birds is well
established (Raphael et al. 1987, Hutto 1995, Saab and Dudley 1998). However, little
information exists for the Sierra Nevada describing the important characteristics in post-fire
snag-dominated habitats that determine the density and diversity of cavity-nesting species. Our
results here provide some of the only detailed information for a whole suite of cavity-nesting
species in post-fire habitats in the Sierra Nevada (also see Raphael and White 1984).
Patterns in nest tree characteristics were remarkably consistent across the three years of
this study. Higher decay classes and broken top snags are still being used more than they are
available. However, in 2011 we started to see more birds using the fire killed snags for nesting
than they did in 2009, suggesting those snags are now becoming suitable for nesting. Blackbacked woodpecker continued to show very little selection for tree species, size, or decay class
through 2011. They readily used what was available on the landscape, suggesting other factors
are driving this species’ occupancy within these fires.
Both Black-backed and Hairy Woodpeckers showed strong selection for high snag
densities surrounding their nest trees. Extrapolating from our 0.1 acre plot, this would suggest
they are selecting for areas with greater than 80 snags per acre or 200 per hectare. This density
coincides well with those reported from other studies in the West (Saab et al. 2009). These other
studies have suggested that temporally limited food resources are the most important factor for
determining these species’ distribution in high severity fire areas (Saab and Dudley 1998, Dixon
and Saab 2000). Thus, management of snag density in post-fire areas should consider
woodpecker nesting and foraging needs.
Private vs. Public Land
Through 2011, the majority of public land had been untouched since these fires, whereas
private land has been extensively salvaged, prepped, and replanted. As such, the differences in
the avian community between the two were considerable. Private land in the Storrie and
Moonlight fires supported a significantly less diverse and abundant avian community than the
surrounding Forest Service land. The reduced avian diversity and overall abundance in salvaged
areas was primarily a result of significantly fewer green forest-associated species on private land.
This is likely a result of severity being higher on private land but also private land management
practices that err on the side of taking trees that might survive versus leaving trees that might die.
This is clearly evidenced at the property boundaries, where no green trees remain on private land
but a substantial number do immediately across the boundary on Forest Service land. This
58
practice results in few if any overstory trees left and thus little habitat for the diverse group of
tree-nesting foliage-gleaning species, especially those associated with green forest edges.
Olive-sided Flycatcher and Western Wood-pewee, two edge species shown to be
declining in the Sierra Nevada (Sauer et al. 2011), were absent from private lands despite being
more abundant in post-fire areas. These birds are not typically associated with one particular
habitat but rather the juxtaposition of unlike habitats. Management practices that produce
homogenous landscapes such as those on private land in the Storrie and Moonlight fires are
unlikely to support these and other species associated with habitat mosaics. Additionally, as
these even-aged stands develop over time they are unlikely to provide habitat for a wide range of
species associated with stand and landscape level heterogeneity. We found several even-aged
plantations that supported significantly lower avian diversity and abundance and nearly lacked
cavity-nesting species 20 to 40 years after planting on the Almanor Ranger District (Burnett et al.
2010). Reducing the threshold from retaining only trees with an 80% or greater chance of
survival to something considerably less would result in greater habitat heterogeneity, a mature
green tree component, and greater snag densities in the long term. All of these aspects would
undoubtedly increase avian diversity in heavily salvaged post-fire landscapes.
In riparian areas on private land avian diversity and abundance were similar to the unsalvaged Forest Service land. This was primarily due to an elevated abundance of understorynesting species. The riparian areas within private land were not entered during salvage operations
or subjected to herbicide treatments in both fires, resulting in far more snags, live trees, shrubs,
and herbaceous vegetation. Leaving more patches of un-salvaged habitat and limiting the use of
herbicides on private land would undoubtedly increase avian diversity in the short term and help
promote a more heterogeneous forest in the future.
Conclusions
There is a growing need to understand the value of the habitats created by wildfire and
the critical elements required by the unique and relatively diverse avian community in the Sierra
Nevada. Wildfires provide a unique opportunity to mold a landscape into the forest composition
that will exist there for decades to come. Though the severity and scale of fires in the Sierra
Nevada in the last few decades may exceed historic averages, there is little doubt that even the
areas that burn at the highest severity support a unique and relatively diverse avian community.
In fact, many of the bird species declining in the Sierra reach their greatest abundance in severely
burned areas, where many habitat elements that have been eliminated or degraded by past
management actions (e.g., snags, herbaceous understory, hardwoods) flourish. The results from
this ongoing study are providing important information to help inform management of burned
areas to ensure wildlife needs are met while addressing other post-fire objectives. It is also
providing valuable information on Black-backed Woodpeckers as they undergo a status review
for consideration to be listed under the California Endangered Species Act. In 2012 we will use
available remotely-sensed data on burn severity, pre-fire habitat composition, and our groundbased habitat data to better understand the importance of severity class, patch size, and snag
densities for the various species associated with post-fire habitat. In addition, we will be
synthesizing results on the effects of fuels treatments in unburned forest and will complete
several publications related to both of these efforts.
59
Literature Cited
Agee, J.K. 1993. Fire ecology of Pacific Northwest Forests. Island Press. Washington, D.C.
Barbour, M.G. Major, J., 1988. Terrestrial Vegetation of California, expanded edition. California
Native Plant Society Special Publication 9. Sacramento,California, pp. 3–10. Summarized
in Stephens,S.L., R.E. Martin, and N.E. Clinton. 2007. Prehistoric fire area and emissions from
California forests, woodlands, shrublands, and grasslands. Forest Ecology and Management 251: 205216.
Brawn, J.D., S.K. Robinson, and F.R. Thompson. 2001. The role of disturbance in the ecology and
conservation of birds. Annual Review of Ecological Systems 32, 251-276.
Burnett, R.D., D. Jongsomjit, and D. Stralberg. 2009. Avian Monitoring in the Lassen and Plumas
National Forests: 2008 Annual Report. PRBO Conservation Science, Petaluma, CA. Contribution
Number 1684.
Burnett, R.D., D.L. Humple, A.M. Fogg, and T. Guida. 2010. Avian Monitoring in the Lassen
National Forest: 2009 Annual Report. PRBO Conservation Science, Petaluma, CA. Contribution
Number 1740.
Dixon, R.D., and V. A. Saab. 2000. Black-backed Woodpecker (Picoides arcticus). In: The Birds of
North America. No. 509. (Poole, A. and Gill, F. eds.). The Academy of Natural Sciences,
Philadelphia, PA, and The American Ornithologists' Union, Washington D.C.
Dudley, J. G., and V. Saab. 2003. A field protocol to monitor cavity-nesting birds. In RMRS-RP-44.
Fort Collins, CO: USDA Forest Service, Rocky Mountain Research Station.
ESRI. 2004. ArcMap [GIS software]. Version 9.2. Redlands, CA: Environmental Systems Research
Institute, 1992-2005.
Foster, D.R., D.H. Knight, and J.F. Franklin. 1998. Landscape patterns and legacies resulting from
large, infrequent forest disturbances. Ecosystems 1:497-510.
Huff, M.H., N.E. Seavy, J.D. Alexander, and C.J. Ralph. 2005. Fire and birds in maritime pacific
northwest. Studies in Avian Biology 30:46-62.
Hutto, R.L. 1995. Composition of bird communities following stand-replacement fires in Northern
Rocky Mountain (U.S.A.) conifer forests. Conservation Biology 9:1041-1058.
Hutto, R.L. 2006. Towards meaningful snag management guidelines for postfire salvage logging in
North American conifer forests. Conservation Biology 20: 984-993.
Key, C. H., and N. C. Benson. 2005. Landscape assessment: ground measure of severity, the
Composite Burn Index. In D. C. Lutes (Ed.), FIREMON: Fire effects monitoring and inventory
system. General Technical Report, RMRS-GTR-164, Ogden, UT: USDA Forest Service, Rocky
Mountain Research Station.
60
Krebs, C.J. 1989. Ecological Methodology. Harper and Row Publishers, New York, New York: 654
pp.
MacArthur, R.H. 1965. Patterns of species diversity. Biological Reviews 40:510-533.
Miller, J.D. and A.E. Thode. 2007. Quantifying burn severity in a heterogenous landscape with a
relative version of the delta Normalized Burn Ratio (dNBR). Remote Sensing of Environment 109:
66-80.
Miller, J.D., H.D. Safford, M. Crimmins, and A.E. Thode. 2008. Quantitative evidence for increasing
forest fire severity in the Sierra Nevada and southern Cascade mountains, California and Nevada,
USA. Ecosystems 12: 16-32.
Nur, N., S.L. Jones, and G.R. Geupel. 1999. Statistical guide to data analysis of avian monitoring
programs. U.S. Fish & Wildlife Service, Biological Technical Publication R6001-1999.
Ralph, C.J., S. Droege, and J.R. Sauer. 1995. Managing and monitoring birds using point counts:
standards and applications. In C. J. Ralph, J. R. Sauer and S. Droege (eds.), Monitoring Bird
Populations by Point Counts. USDA Forest Service Publication, Gen. Tech. Rep. PSW-GTR-149,
Albany, CA .
Raphael, M. G., and M. White. 1984. Use of snags by cavity-nesting birds in the Sierra-Nevada.
Wildlife Monographs 86:1-66.
Raphael, M.G., M.L. Morrison, and M.P. Yoder-Williams. 1987. Breeding bird populations during
twenty five years of post-fire succession in the Sierra Nevada. The Condor 89:614-626.
Reynolds, R.T., J.M. Scott, and R.A. Nussbaum. 1980. A variable circular plot method for estimating
bird numbers. Condor 82:309:313.
Saab, V. A. and J.G. Dudley. 1998. Responses of cavity-nesting birds to stand replacement fire and
salvage logging in ponderosa pine/Douglas-fir forests of southwestern Idaho. Res. Paper RMRS-RP11. Ogden, UT: Rocky Mountain Research Station, Forest Service, U.S. Department of Agriculture;
17 p.
Saab, V., A. Brannon, J. Dudley, L. Donohoo, D. Vanderzanden, V. Johnson, and H. Lachowski.
2002. Selection of fire-created snags at two spatial scales by cavity-nesting birds. Gen. Tech. Rep.
PSW-GTR-181. http://www.fs.fed.us/psw/publications/documents/gtr-181/062_Saab.pdf
Saab, V.A., J. Dudley, and W. Thompson. 2004. Factors influencing occupancy of nest cavities in
recently burned forests. The Condor 106:20-36.
Saab, V.A., R.E. Russell, and J.G. Dudley. 2007. Nest densities of cavity-nesting birds in relation to
postfire salvage logging and time since wildfire. The Condor 109:97-108.
61
Saab, V.A., R.E. Russell, and J. Dudley. 2009. Nest-site selection by cavity-nesting birds in relation to
postfire salvage logging. Forest Ecology and Management 257:151-159.
Sauer, J. R., J. E. Hines, J. E. Fallon, K. L. Pardieck, D. J. Ziolkowski, Jr., and W. A. Link. 2011. The
North American Breeding Bird Survey, Results and Analysis 1966 - 2010. Version 12.07.2011 USGS
Patuxent Wildlife Research Center, Laurel, MD
Siegel, R.B., J.F. Saracco, and R.L. Wilkerson. 2010. Management Indicator Species (MIS) Surveys
on Sierra Nevada National Forests: Black-backed Woodpecker: 2009 Annual Report. The Institute for
Bird Populations, Point Reyes Station,
CA. http://www.birdpop.org/DownloadDocuments/2009_MIS_report_for_BBWO.pdf
Smucker, K.M., R.L. Hutto, and B.M. Steele. 2005. Changes in bird abundance after wildfire:
importance of fire severity and time since fire. Ecological Applications 15:1535-1549.
Stata Corp. 2007. Stata/IC 10.0 for Windows. StataCorp LP. College Station, TX. USA
Stine, P., M. Landram, J. Keane, D. Lee, B. Laudenslayer, P. Weatherspoon, and J. Baldwin. 2002.
Fire and fuels management, landscape dynamics, and fish and wildlife resources: An integrated
research plan for the Herger-Feinstein Quincy Library Group Pilot Study
Area. http://www.fs.fed.us/psw/programs/snrc/forest_health/plas_studyplan.pdf
62
Appendix A. Models of the effect of burn severity on avian community
metrics.
Burn Severity Effect on Shannon Index of Avian Diversity
All Fires – Linear
Source |
SS
df
MS
Number of obs = 238
-------------+-----------------------------F( 1, 236) = 0.00
Model | .039
1
.04
Prob > F = 0.9497
Residual | 2329.64 236 9.87
R-squared = 0.0000
-------------+-----------------------------Adj R-squared = -0.0042
Total | 2329.68 237 9.83
Root MSE = 3.1419
-----------------------------------------------------------------------------sw | Coef.
Std. Err.
t
P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | -.0129807 .2056272 -0.06 0.950 -.41808 .3921186
cons | 9.201774 .3728396 24.68 0.000 8.467255 9.936293
------------------------------------------------------------------------------
All Fires – Quadratic
Source |
SS
df
MS
Number of obs = 238
-------------+-----------------------------F( 2, 235) = 8.11
Model | 150.41 2 75.21
Prob > F = 0.0004
Residual | 2179.27 235 9.27
R-squared = 0.0646
-------------+-----------------------------Adj R-squared = 0.0566
Total | 2329.68 237 9.83
Root MSE = 3.0452
-----------------------------------------------------------------------------sw | Coef.
Std. Err.
t
P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | -.1876925 .2039705 -0.92 0.358 -.5895368 .2141518
cbistd2 | -.9923164 .2464236 -4.03 0.000 -1.477798 -.5068347
cons | 10.45303 .476593 21.93 0.000 9.51409 11.39197
-----------------------------------------------------------------------------Test for Goodness of Fit of Quadratic over Linear
Likelihood-ratio test
(Assumption: a nested in .)
LR chi2(1) = 15.88
Prob > chi2 = 0.0001
63
Cub Fire - Linear
Source |
SS
df
MS
Number of obs = 54
-------------+-----------------------------F( 1, 52) = 10.25
Model | 85.744483 1 85.74
Prob > F = 0.0023
Residual | 434.882529 52 8.36
R-squared = 0.1647
-------------+-----------------------------Adj R-squared = 0.1486
Total | 520.627012 53 9.82
Root MSE = 2.8919
-----------------------------------------------------------------------------sw | Coef.
Std. Err.
t P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi |1.845325 .5763072 3.20 0.002 .6888798 3.00177
cons |7.216832 .6688934 10.79 0.000 5.874599 8.559065
------------------------------------------------------------------------------
Cub Fire - Quadratic
Source |
SS
df
MS
Number of obs = 54
-------------+-----------------------------F( 2, 51) = 6.95
Model | 111.566824 2 55.78
Prob > F = 0.0021
Residual | 409.060188 51 8.02
R-squared = 0.2143
-------------+-----------------------------Adj R-squared = 0.1835
Total | 520.627012 53 9.82
Root MSE = 2.8321
-----------------------------------------------------------------------------sw | Coef. Std. Err.
t
P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | 1.087257 .705006 1.54 0.129 -.3281033 2.502616
cbistd2 | -1.18411 .659937 -1.79 0.079 -2.508991 .1407697
cons | 9.048374 1.21287 7.46 0.000 6.61342 11.48333
-----------------------------------------------------------------------------Test for Goodness of Fit of Quadratic over Linear
Likelihood-ratio test
(Assumption: a nested in .)
LR chi2(1) = 3.31
Prob > chi2 = 0.0690
64
Moonlight Fire – Linear
Source |
SS
df
MS
Number of obs = 129
-------------+-----------------------------F( 1, 127) = 3.64
Model | 36.72 1 36.72
Prob > F = 0.0587
Residual |1281.96 127 10.09
R-squared = 0.0278
-------------+-----------------------------Adj R-squared = 0.0202
Total | 1318.68 128 10.30
Root MSE = 3.1771
-----------------------------------------------------------------------------sw | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | -.5408663 .2835703 -1.91 0.059 -1.102001 .0202681
cons | 10.26458 .5677839 18.08 0.000 9.141036 11.38812
------------------------------------------------------------------------------
Moonlight Fire – Quadratic
Source |
SS
df
MS
Number of obs = 129
-------------+-----------------------------F( 2, 126) = 4.32
Model | 84.53
2 42.26
Prob > F = 0.0154
Residual | 1234.15 126 9.79
R-squared = 0.0641
-------------+-----------------------------Adj R-squared = 0.0493
Total | 1318.69 128 10.30
Root MSE = 3.1
-----------------------------------------------------------------------------sw | Coef.
Std. Err.
t
P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | -.5806961
.2799145 -2.07 0.040 -1.134639 -.0267534
cbistd2 | -.7303744 .3305633 -2.21 0.029 -1.384549 -.0761993
cons | 11.05376 .6636233 16.66 0.000
9.740474 12.36706
-----------------------------------------------------------------------------Test for Goodness of Fit of Quadratic over Linear
Likelihood-ratio test
(Assumption: a nested in .)
LR chi2(1) = 4.90
Prob > chi2 = 0.0268
Storrie fire – Linear
65
. fit sw cbi if fire=="ST" & private==0
Source |
SS
df
MS
Number of obs = 55
-------------+-----------------------------F( 1, 53) = 0.10
Model | .91
1
.92
Prob > F = 0.7524
Residual | 483.43 53 9.12
R-squared = 0.0019
-------------+-----------------------------Adj R-squared = -0.0169
Total | 484.35
54 8.97
Root MSE = 3.0202
-----------------------------------------------------------------------------sw | Coef.
Std. Err.
t P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | .1257289 .3965484
0.32 0.752 -.6696464 .9211041
_cons | 8.885906 .7419523 11.98 0.000 7.397738 10.37407
-----------------------------------------------------------------------------Storrie fire – Quadratic
Source |
SS
df
MS
Number of obs = 55
-------------+-----------------------------F( 2, 52) = 2.17
Model | 37.36 2 18.68
Prob > F = 0.1241
Residual | 446.99 52 8.60
R-squared = 0.0771
-------------+-----------------------------Adj R-squared = 0.0416
Total | 484.35 54 8.97
Root MSE = 2.9319
-----------------------------------------------------------------------------sw | Coef.
Std. Err.
t
P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | -.2221794 .4204127 -0.53 0.599 -1.065799 .6214403
cbistd2 | -1.031637 .5010589 -2.06 0.045 -2.037085 -.0261883
_cons | 10.52268 1.072738
9.81 0.000 8.370076 12.67529
-----------------------------------------------------------------------------Test for Goodness of Fit of Quadratic over Linear
Likelihood-ratio test
(Assumption: a nested in .)
LR chi2(1) = 4.31
Prob > chi2 = 0.0379
66
Burn Severity Effect on Total Bird Abundance
All Fires – Linear
Source |
SS
df
MS
Number of obs = 238
-------------+-----------------------------F( 1, 236) = 0.00
Model | .01
1
.01
Prob > F = 0.9773
Residual | 1531.24 236 6.49
R-squared = 0.0000
-------------+-----------------------------Adj R-squared = -0.0042
Total | 1531.24 237 6.46
Root MSE = 2.5472
-----------------------------------------------------------------------------indpervis | Coef. Std. Err.
t P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | .0047399 .1667082 0.03 0.977 -.3236863 .3331662
cons | 5.696583 .3022723 18.85 0.000 5.101086 6.292079
-----------------------------------------------------------------------------All Fires – Quadratic
Source |
SS
df
MS
Number of obs = 238
-------------+-----------------------------F( 2, 235) = 2.90
Model | 36.86
2 18.43
Prob > F = 0.0571
Residual |1494.38 235 6.36
R-squared = 0.0241
-------------+-----------------------------Adj R-squared = 0.0158
Total | 1531.24 237 6.46
Root MSE = 2.5217
-----------------------------------------------------------------------------indpervis | Coef. Std. Err.
t
P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | -.0817539 .168905 -0.48 0.629 -.4145154 .2510076
cbistd2 | -.4912619 .2040599 -2.41 0.017 -.8932824 -.0892415
cons | 6.316037 .3946599 16.00 0.000 5.538514 7.09356
-----------------------------------------------------------------------------Test for Goodness of Fit of Quadratic over Linear
Likelihood-ratio test
(Assumption: a nested in .)
LR chi2(1) = 5.80
Prob > chi2 = 0.0160
Cub Fire - Linear
Source |
SS
df
MS
Number of obs =
54
67
-------------+-----------------------------F( 1, 52) = 0.57
Model | 4.32
1 4.32
Prob > F = 0.4544
Residual | 395.73 52 7.61
R-squared = 0.0108
-------------+-----------------------------Adj R-squared = -0.0082
Total | 400.05
53 7.55
Root MSE = 2.7586
-----------------------------------------------------------------------------indpervis | Coef. Std. Err.
t
P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | .4143932 .5497508 0.75 0.454 -.6887624 1.517549
cons | 4.824047 .6380706 7.56 0.000 3.543665 6.10443
------------------------------------------------------------------------------
Cub Fire - Quadratic
Source |
SS
df
MS
Number of obs = 54
-------------+-----------------------------F( 2, 51) = 0.29
Model | 4.52
2 2.25808164
Prob > F = 0.7486
Residual | 395.53 51 7.75558358
R-squared = 0.0113
-------------+-----------------------------Adj R-squared = -0.0275
Total | 400.05
53 7.55
Root MSE = 2.7849
-----------------------------------------------------------------------------indpervis | Coef. Std. Err.
t P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | .3489983 .6932533 0.50 0.617 -1.042766 1.740762
cbistd2 | -.1021475 .6489357 -0.16 0.876 -1.40494 1.200645
_cons | 4.982046 1.192658 4.18 0.000
2.587685 7.376407
-----------------------------------------------------------------------------Test for Goodness of Fit of Quadratic over Linear
Likelihood-ratio test
(Assumption: a nested in .)
LR chi2(1) = 0.03
Prob > chi2 = 0.8713
Moonlight Fire – Linear
Source |
SS
df
MS
-------------+------------------------------
Number of obs = 129
F( 1, 127) = 8.45
68
Model | 56.21
1 56.21
Prob > F = 0.0043
Residual | 844.43 127 6.65
R-squared = 0.0624
-------------+-----------------------------Adj R-squared = 0.0550
Total | 900.63 128 7.04
Root MSE = 2.5786
-----------------------------------------------------------------------------indpervis | Coef.
Std. Err.
t
P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | -.6691473 .2301459 -2.91 0.004 -1.124564 -.2137302
cons | 7.008948 .4608138 15.21 0.000 6.09708 7.920815
------------------------------------------------------------------------------
Moonlight Fire – Quadratic
Source |
SS
df
MS
-------------+-----------------------------Model | 74.56
2 37.28
Residual | 826.07 126 6.56
-------------+-----------------------------Total | 900.63
128 7.04
Number of obs = 129
F( 2, 126) = 5.69
Prob > F = 0.0043
R-squared = 0.0828
Adj R-squared = 0.0682
Root MSE = 2.5605
-----------------------------------------------------------------------------indpervis | Coef. Std. Err.
t
P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | -.6938253 .2290076 -3.03 0.003 -1.147025 -.2406261
cbistd2| -.4525304 .2704451 -1.67 0.097 -.9877333 .0826724
cons | 7.497918 .5429327 13.81 0.000 6.42347 8.572366
-----------------------------------------------------------------------------Test for Goodness of Fit of Quadratic over Linear
Likelihood-ratio test
(Assumption: a nested in .)
LR chi2(1) = 2.84
Prob > chi2 = 0.0922
Storrie fire – Linear
Source |
SS
df
MS
-------------+-----------------------------Model | 47.52 1 47.53
Number of obs = 55
F( 1, 53) = 8.82
Prob > F = 0.0045
69
Residual |285.45 53 5.39
R-squared = 0.1427
-------------+-----------------------------Adj R-squared = 0.1265
Total | 332.97 54 6.17
Root MSE = 2.3207
-----------------------------------------------------------------------------indpervis | Coef.
Std. Err.
t P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | .9051152 .3047148 2.97 0.004 .2939347 1.516296
cons | 4.543491 .5701294 7.97 0.000 3.399956 5.687025
------------------------------------------------------------------------------
Storrie fire – Quadratic
Source |
SS
df
MS
Number of obs = 55
-------------+-----------------------------F( 2, 52) = 5.50
Model | 58.16
2 29.08
Prob > F = 0.0068
Residual | 274.81 52 5.28
R-squared = 0.1747
-------------+-----------------------------Adj R-squared = 0.1429
Total | 332.97 54 6.17
Root MSE = 2.2989
-----------------------------------------------------------------------------indpervis | Coef. Std. Err.
t
P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | .7170877 .3296395 2.18 0.034 .0556177 1.378558
cbistd2 | -.5575495 .3928731 -1.42 0.162 -1.345907 .2308081
cons | 5.428089 .8411183 6.45 0.000 3.740261 7.115916
-----------------------------------------------------------------------------Test for Goodness of Fit of Quadratic over Linear
Likelihood-ratio test
(Assumption: a nested in .)
LR chi2(1) = 2.09
Prob > chi2 = 0.1483
Private Land Burn Severity Effect on Shannon Index of Avian Diversity
Storrie fire – Linear
Source |
SS
df
MS
Number of obs =
20
70
-------------+-----------------------------F( 1, 18) = 1.68
Model | 10.28 1 10.28
Prob > F = 0.2110
Residual | 110.00 18 6.11
R-squared = 0.0855
-------------+-----------------------------Adj R-squared = 0.0347
Total | 120.28 19 6.33
Root MSE = 2.4721
-----------------------------------------------------------------------------indpervis | Coef.
Std. Err.
t
P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | 1.544273 1.190498 1.30 0.211 -.9568703 4.045416
cons | .216883 3.091542 0.07 0.945 -6.278206 6.711972
------------------------------------------------------------------------------
Storrie fire – Quadratic
Source |
SS
df
MS
Number of obs = 20
-------------+-----------------------------F( 2, 17) = 2.07
Model | 23.59 2 11.79
Prob > F = 0.1564
Residual | 96.70 17 5.69
R-squared = 0.1961
-------------+-----------------------------Adj R-squared = 0.1015
Total | 120.28 19 6.33
Root MSE = 2.385
-----------------------------------------------------------------------------indpervis | Coef.
Std. Err.
t
P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | -1.102668 2.077185 -0.53 0.602 -5.485144 3.279808
cbistd2 | 2.37748 1.554568 1.53 0.145 -.9023727 5.657333
cons | 4.437707 4.063601 1.09 0.290 -4.135742 13.01116
-----------------------------------------------------------------------------Test for Goodness of Fit of Quadratic over Linear
Likelihood-ratio test
(Assumption: a nested in .)
LR chi2(1) = 2.58
Prob > chi2 = 0.1084
Moonlight Fire - Linear
Source |
SS
df
MS
-------------+------------------------------
Number of obs = 30
F( 1, 28) = 4.12
71
Model | 9.36
1 9.36
Prob > F = 0.0520
Residual | 63.61 28 2.27
R-squared = 0.1282
-------------+-----------------------------Adj R-squared = 0.0971
Total | 72.96
29 2.52
Root MSE = 1.5072
-----------------------------------------------------------------------------indpervis | Coef.
Std. Err.
t P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | -.5882569 .2898696 -2.03 0.052 -1.182028 .0055141
cons | 5.241523 .7579351 6.92 0.000 3.688964 6.794083
------------------------------------------------------------------------------
Moonlight Fire – Quadratic
Source |
SS
df
MS
Number of obs = 30
-------------+-----------------------------F( 2, 27) = 2.20
Model | 10.23
2 5.11
Prob > F = 0.1302
Residual | 62.73 27 2.32
R-squared = 0.1402
-------------+-----------------------------Adj R-squared = 0.0765
Total | 72.96
29 2.52
Root MSE = 1.5243
-----------------------------------------------------------------------------indpervis | Coef. Std. Err.
t P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | -.6146719 .296324 -2.07 0.048 -1.222678 -.0066653
cbistd2 | .2866468 .4684363 0.61 0.546 -.6745051 1.247799
cons | 4.861685 .9863563 4.93 0.000 2.837849 6.88552
-----------------------------------------------------------------------------Test for Goodness of Fit of Quadratic over Linear
Likelihood-ratio test
(Assumption: a nested in .)
LR chi2(1) = 0.41
Prob > chi2 = 0.5204
Private Land Burn Severity Effect on Total Bird Abundance
Storrie fire – Linear
72
Source |
SS
df
MS
Number of obs = 20
-------------+-----------------------------F( 1, 18) = 1.68
Model | 10.28
1 10.28
Prob > F = 0.2110
Residual | 110.00 18 6.11
R-squared = 0.0855
-------------+-----------------------------Adj R-squared = 0.0347
Total | 120.28
19 6.33
Root MSE = 2.4721
-----------------------------------------------------------------------------indpervis | Coef. Std. Err.
t P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | 1.544273 1.190498 1.30 0.211 -.9568703 4.045416
cons | .216883 3.091542 0.07 0.945 -6.278206 6.711972
------------------------------------------------------------------------------
Storrie fire – Quadratic
Source |
SS
df
MS
Number of obs = 20
-------------+-----------------------------F( 2, 17) = 2.07
Model | 23.59 2 11.79
Prob > F = 0.1564
Residual | 96.69 17 5.69
R-squared = 0.1961
-------------+-----------------------------Adj R-squared = 0.1015
Total | 120.28 19
6.33
Root MSE = 2.385
-----------------------------------------------------------------------------indpervis | Coef. Std. Err.
t P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | -1.102668 2.077185 -0.53 0.602 -5.485144 3.279808
cbistd2 | 2.37748 1.554568 1.53 0.145 -.9023727 5.657333
cons | 4.437707 4.063601 1.09 0.290 -4.135742 13.01116
-----------------------------------------------------------------------------Test for Goodness of Fit of Quadratic over Linear
Likelihood-ratio test
(Assumption: a nested in .)
LR chi2(1) = 2.58
Prob > chi2 = 0.1084
Moonlight – Linear
Source |
SS
df
MS
-------------+------------------------------
Number of obs = 30
F( 1, 28) = 4.12
73
Model | 9.35
1 9.36
Prob > F = 0.0520
Residual | 63.61
28 2.27
R-squared = 0.1282
-------------+-----------------------------Adj R-squared = 0.0971
Total | 72.96
29
2.52
Root MSE = 1.5072
-----------------------------------------------------------------------------indpervis | Coef. Std. Err.
t
P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | -.5882569 .2898696 -2.03 0.052 -1.182028 .0055141
cons | 5.241523 .7579351 6.92 0.000 3.688964 6.794083
------------------------------------------------------------------------------
Moonlight - Quadratic
Source |
SS
df
MS
Number of obs = 30
-------------+-----------------------------F( 2, 27) = 2.20
Model | 10.23
2 5.11
Prob > F = 0.1302
Residual | 62.73 27 2.32
R-squared = 0.1402
-------------+-----------------------------Adj R-squared = 0.0765
Total | 72.96
29 2.52
Root MSE = 1.5243
-----------------------------------------------------------------------------indpervis | Coef.
Std. Err.
t P>|t| [95% Conf. Interval]
-------------+---------------------------------------------------------------cbi | -.6146719 .296324 -2.07 0.048 -1.222678 -.0066653
cbistd2 | .2866468 .4684363 0.61 0.546 -.6745051 1.247799
cons | 4.861685 .9863563 4.93 0.000 2.837849 6.88552
-----------------------------------------------------------------------------Test for Goodness of Fit of Quadratic over Linear
Likelihood-ratio test
(Assumption: a nested in .)
LR chi2(1) = 0.41
Prob > chi2 = 0.5204
74
Chapter 3:
Spotted Owl Module
Dr. John Keane, Dr. Mary Conner, Claire V. Gallagher, Dr. Ross A. Gerrard,
Gretchen Jehle, Paula A. Shaklee
Executive Summary:
Our efforts from 2003-2011 have focused on monitoring California spotted owl (CSO) (Strix
occidentalis occidentalis) distribution, abundance, and demographics to address our primary
research objectives and provide the baseline data for assessing the effects of HFQLG
implementation. In conjunction with the now fully integrated Lassen Demographic Study, we
have collected landscape-scale information on the distribution and abundance of CSOs
across approximately 650,000 acres of land. Determining the accurate number and
distribution of CSO sites requires multiple years of survey and marking of individual CSOs
to delineate separate territories and identify individual birds that move among multiple sites
within and across years. These baseline data are fundamental for developing empiricallybased habitat models for understanding CSO habitat associations and developing adaptive
management tools and models. The completion of the Meadow Valley area projects in 20072008 marked the first landscape series of HFQLG treatments to be implemented within the
study area, providing the first opportunity to address treatment effects within a case study.
Additionally, we assessed CSO distribution and abundance in response to high-intensity
(Moonlight-Antelope Complex) and low/moderate-intensity (Cub-Onion Complex) wildfires
during 2008-2010. Our baseline information on landscape CSO distribution and habitat
associations, coupled with our 2007-2008 radio- telemetry work, will allow us to assess
associations between CSOs and vegetation changes resulting from fuels and forest
management treatments and from wildfire.
In 2011 we completed our fourth year of monitoring CSO distribution and abundance in the
Meadow Valley Project Area. During 2008-2010 we observed similar numbers of territorial
owl sites (6-7) in the post-treatment landscape, although we did observe territory-specific
responses in both areas that have experienced treatments and sites where little treatment has
occurred. In 2011 we documented 4 territories within the Meadow Valley Project Area. This
was a reduction of 3 territorial sites compared to 2010, whereas we had a net reduction of 2
additional territorial sites across the rest of the PLS area. Our assessment of CSO response to
wildfire suggests that CSO are able to persist in landscapes that experience primarily
low/moderate severity wildfire, whereas landscapes that experience primarily high-severity
do not support comparable numbers or distribution of CSOs. Our radio-telemetry worked
completed in 2010 has provided insight into how individual owls respond to treatments
within the first 1-2 years following implementation. Results suggest that CSOs avoid use of
Defensible Fuel Profile Zone treatments.
We completed analysis of the demographic data from the 20-year long demographic study of
CSOs on the Lassen Demographic Study Area embedded within our overall study. These
results suggest that CSOs on the study area have experienced a long-term population decline
with an estimated annual population growth rate (lambda) of 0.986 (SE = 0.010) with 95%
75
confidence intervals ranging of 0.967-1.006. Our biggest challenge to date has been the lack
of accurate vegetation information to develop predictive habitat models and to document
changes in vegetation due to treatment or disturbance. In 2010 we contracted to obtain
vegetation maps that will support development of predictive habitat models and be sufficient
to document changes in vegetation over space and time that can be used to understand CSO
response to vegetation change. In summary, we are working towards being able to broadly
address CSO management questions across a gradient of landscape conditions ranging across
untreated landscapes, landscapes treated to meet desired fuels/vegetation conditions, and
landscapes that have experienced wildfire in order to address primary management issues.
Introduction
Knowledge regarding the effects of fuels and vegetation management on California spotted
owls (Strix occidentalis occidentalis; CSOs) and their habitat is a primary information need
for addressing conservation and management objectives in Sierra Nevada forests. The
specific research objectives of the California spotted owl module as identified and described
in the Plumas-Lassen Study (PLS) Plan are:
1) What are the associations among landscape fuels treatments and CSO density, distribution,
population trends and habitat suitability at the landscape-scale?
2) What are the associations among landscape fuels treatments and CSO reproduction,
survival, and habitat fitness potential at the core area/home range scales?
3) What are the associations among landscape fuels treatments and CSO habitat use and
home range configuration at the core area/home range scale?
4) What is the population trend of CSO in the northern Sierra Nevada, and which factors
account for variation in population trend?
5) Are barred owls increasing in the northern Sierra Nevada, what factors are associated with
their distribution and abundance, and are they associated with reduced CSO territory
occupancy?
6) Does West Nile Virus affect the survival, distribution and abundance of California spotted
owls in the study area?
7) What are the effects of wildfire on California spotted owls and their habitat?
Our focus in 2011 was to conduct landscape inventories of CSO distribution and abundance
and continue banding to provide the required data and baseline information to meet the
objectives of Research Questions 1-4 identified above. Current information on the
distribution and density of CSOs across the HFQLG study area is required to provide the
data necessary to build predictive habitat models and provide baseline population
information against which we will assess post-treatment changes in CSO populations and
habitat. Continued monitoring on the Lassen Demographic Study Area is critical for
estimating CSO population trends and status. Complete landscape inventory surveys were
76
conducted across 9 of the 11 original project survey areas in 2011 (Figure 1). Surveys were
not conducted in 2 survey areas in 2006-2011 (SA-5, SA-7, Figure 1); sufficient data for
determining the number and distribution of CSO sites for initial habitat modeling efforts was
collected in these areas from 2004-2005. In 2011, we conducted the first year of surveys
within the Storrie Fire footprint in priority areas designated by Plumas National Forest staff.
Details on survey methods are described in the study plan. Efforts were made to monitor the
pair and reproductive status of each owl and to capture, uniquely color-mark, and collect
blood samples from each individual owl across the study area. Capture and color-marking
are necessary to estimate survival and population trend and to assess exposure to West Nile
Virus (WNV) (Research Question #5). We also recorded all barred and hybrid barred-spotted
owls encountered in the study area and synthesized all existing barred owl records for the
northern Sierra Nevada to address Research Question #6. Additionally, we completed a
radio-telemetry study on CSOs within SA-4 in the Meadow Valley project area to document
home range size and configuration and to assess habitat associations relative to the recently
implemented treatments. In response to a need for information on the association between
CSOs and wildfire, we conducted a second year of surveys to assess CSO distribution,
abundance, and habitat associations in the Cub-Onion Complex fire area (COCFA) on the
Lassen NF. The information from the COCFA complements our data from the MoonlightAntelope Complex fire area (MACFA) collected in 2008-2009. The MACFA fires burned in
2007 and we conducted surveys in 2008 and 2009 to assess the immediate post-fire response
of CSOs. The COCFA burned in 2008 and we conducted surveys in 2009 and 2010. For both
study areas, we surveyed for CSOs in the first and second years immediately post-fire.
Finally, we completed an analysis of the demographic data from the Lassen Demographic
Study area collected through 2011 to provide the most current information on the status,
demography and population trend of CSOs in the northern Sierra Nevada.
Results
CSO Numbers, Reproductive Success, Density and Population Trends
A total of 7,215 surveys were conducted over 1,889 km2 in 2011. Sixty-six territorial CSO
sites were documented across the core PLS study area in 2011 (Figure 2). This total consisted
of 48 confirmed pairs, 2 unconfirmed pairs (i.e., one member of pair confirmed as territorial
single plus single detection of opposite sex bird), and 16 territorial single CSOs (single owl
detected multiple times with no pair-mate detected). Only one pair successfully reproduced
in 2011 (2.0% of confirmed/unconfirmed pairs). The single successful nest was located on
the Lassen Demography Study Area and fledged 2 young (Table 1). Across the last eight
years of the study, CSO reproduction has been highest in 2004, 2007, 2009 and 2010 in
terms of the percent of CSO pairs that successfully reproduced and in terms of the number of
young fledged per successful nest. Approximately 50% of CSO pairs successfully
reproduced in 2004, 2007, 2009 and 2010, whereas the proportion of pairs successfully
reproducing ranged between 14% and 18% in 2005, 2006 and 2008. The number of young
produced per successful nest was more similar across years, ranging from 1.47 to 1.81 among
years. CSO reproduction in 2011 was the lowest yet recorded.
77
CSO reproduction is known to vary with spring weather; precipitation patterns were more
similar in 2004 and 2007, with total precipitation relatively low during March-April of 2004
and 2007 as compared to 2005 and 2006 (Figure 3). From 2004 through 2007, CSO
reproduction was high in years of low spring precipitation and low in years with high spring
precipitation. However, this pattern between spring precipitation and reproduction varied
during 2008-2010. In 2008 spring precipitation was low in March-April, yet CSO
reproduction was also low. In contrast, during 2009 spring precipitation was high in
February-March and low in April, and CSO reproduction was high on the Lassen portion of
the study area and low on the Plumas portion of the study area. In 2010, precipitation was
moderate and consistent over the February-May period and CSO reproduction was high.
During 2011 precipitation was high while CSO reproduction was the lowest recorded on the
study site. Rain and snow continued to fall during May and into June during 2011, with a
deep snowpack lingering into mid-summer. It is likely that the heavy early and late spring
precipitation was a primary factor in the historically low CSO reproduction observed in
2011. Results across the years indicate that additional factors influence CSO energetics and
are associated with annual variation in CSO reproduction. Potential factors include
elevational variation in cold and hot temperatures, precipitation, duration of spring/summer,
and snowpack, in addition to annual variation in prey populations.
The Lassen Demographic Study Area (SA-1A, SA-11, SA-12, SA-13, SA-14, SA-15) and
Plumas NF Survey Areas (SA-2, SA-3, SA-4, SA-5, SA-7) were fully integrated in 2005 to
define the overall Plumas-Lassen Study project area and provide consistent CSO survey
effort across the project area (Figures 1 & 2). We estimated the crude density of CSOs based
on the number of territorial owls detected across 9 survey areas during 2011 at the Survey
Area spatial scales (Tables 2 and 3). The estimated crude density across the overall study
area in 2011 was 0.059 territorial owls/km2. Overall study area crude densities are not
directly comparable across years because different total areas were surveyed in each year.
The number of territorial sites (pair sites plus sites with territorial singles) across individual
Survey Areas on the Plumas NF (SA-2, SA-3, SA-4) and Lassen NF (SA-1A, SA-12, SA-13,
SA-14, SA-15) that were surveyed consistently each year between 2005 and 2007 ranged
between a high of 76 in 2004 to a low of 61 in 2005 (Table 3). In the most recent three years
the numbers have ranged from 64 in 2009 to 71 in 2010 and 64 in 2011.
We conducted a 5-year update and re-analyzed the demographic data through 2011 following
the methods described in Blakesley et al. (2010). These data continue to provide the best
estimates of CSO population trends. The Lassen Demographic Study Area is contained
within the overall PLS study area and consists of survey areas SA-1A, SA-11, SA-12, SA13, SA-14 and SA-15 in Figure 1. The estimated mean lambda for the Lassen Demographic
Study between 1990 and 2011 was 0.986 (SE = 0.010), with 95% confidence limits ranging
from 0.967 to 1.006. There was no evidence of linear, quadratic or pseudo-threshold trends
in lambda; rather, the means model was strongly supported by the data. These results suggest
a long-term decline in the CSO population within the Lassen study area over the 20-year
study period. Annual lambda estimates from the best model ranged between 0.87 and 1.13.
Estimates of realized population change based on the time series of lambda estimates
generated from our modeling further suggests that there have been declines in the number of
territory-holding CSOs within the study area (Figure 4). These updated results are similar to
the findings reported for the 1990-2005 period reported in Blakesley et al. (2010).
78
Habitat Assessment – Nest/Roost Plot Scale
We documented a total of 103 CSO territorial sites between 2004 and 2006. We overlayed
the primary nest/roost locations for each of the 103 CSO sites with the CWHR vegetation
classes available within the VESTRA photo-interpreted vegetation map for the PLS to
examine nest/roost-site habitat association patterns. Approximately 53% of the nest sites
were located within CWHR 5M, 5D and 6 size classes (Table 4, Figure 5). An additional
37% of the sites were located within CWHR size class 4M and 4D polygons. CWHR size
class 4 is defined as stands with average tree sizes of 12-24 inch diameter-at-breast-height
(dbh) trees. Of the 38 sites located in size class 4 polygons, 25 (66%) were in size class 4
polygons with a large tree component (i.e., presence of >24 inch dbh trees). Overall, about
90% of the sites were located within CWHR 4M, 4D, 5M, 5D, and 6 size classes. The
remaining 10 sites were located in more open, smaller-tree size polygons, with nests or
roosts located within remnant, scattered larger trees (Table 4, Figure 5).
While the distribution of nest site locations relative to broad vegetation classes provides
insight into patterns of nest-site habitat, we also conducted vegetation sampling at nest or
primary roost sites to describe vegetation structure and composition. Vegetation plot
sampling was conducted at 80 CSO territories across 2005-2007. Vegetation plots were
centered on CSO nest trees, or on a primary roost tree for sites where no nest has been
documented, and were measured using the national Forest and Inventory Assessment (FIA)
protocol. The FIA protocol is used nationally by the USDA Forest Service for inventorying
and monitoring vegetation. FIA sampling consists of measuring vegetation structural and
compositional variables within a 1-ha plot centered on a CSO nest or roost tree. Only one
plot was collected from each CSO territory, with the most frequently used nest tree serving
as the plot center location or the most recent nest tree used at sites where no nest tree was
used more frequently than another. CSO nest sites were characterized by mean total basal
areas of 260.8 ft2/acre, 7.4 snags (>15 inch dbh)/acre, and 10.7 trees (>30 inch dbh)/acre
(Table 5). Under the FIA protocol, canopy cover is modeled based on the tree inventory list.
The modeled canopy cover for these plots averaged 64.1%. Shrub cover averaged 7.7%. Fuel
loads averaged 0.75 tons/acre for 1-hr fuels, 4.0 tons/acre for 10-hr fuels, and 4.44 tons/acre
for 100-hr fuels (Table 5). Use of the FIA sampling protocol will facilitate monitoring of
vegetation and development of CSO habitat models that can be used as adaptive management
planning tools. Habitat models are currently being evaluated that can be used to assess
projected changes in CSO nesting habitat suitability under varying fuels and vegetation
treatment scenarios.
Habitat Assessment – Core Area/Home Range Scale
Core area habitat associations around 102 CSO nest and roost sites was assessed by using a
Geographic Information System (GIS) and the VESTRA photo-interpreted vegetation map to
determine the vegetation patterns within a 500 acre (201 ha) circle centered on each of the
CSO territory sites. To compare the CSO sites with the general availability of habitat across
the study area, we also assessed the same vegetation patterns around 130 points determined
by placing a systematic grid across the study area. For this summary we assessed vegetation
using the USDA Forest Service Region 5 classification system. Overall, CSO core areas
79
averaged 75.7% suitable habitat (classes 3N, 3G, 4N, 4G), whereas the grid points averaged
61.9% (Table 6, Figure 6). Approximately 32% of CSO core areas was composed of large
tree polygons (>24inch dbh, >=40% canopy cover) compared to 19.6% of the grid points
(Table 6, Figure 7).
Meadow Valley Project Area Case Study
The Meadow Valley Project Area (MVPA) is the first area within the PLS where the full
implementation of HFQLG treatments has occurred. Treatments were implemented on the
ground within this project area during 2001-2008, with primarily light thinning and underburning occurring in 2001-2005 and Defensible Fuel Profile Zones and Group Selections
implemented during 2005-2008. The MVPA corresponds closely with the boundaries of SA4 of the PLS (Figure 8).
We began monitoring CSOs in SA-4 in 2003 and have annually monitored the distribution,
abundance and reproduction of CSOs within SA-4. Additionally, we have color-banded all
individuals within this area, with the exception of one male who could not be captured. Full
survey methods are described in detail in our study plan (available from field project leaders)
and are consistent with USDA Forest Service R5 survey methods. Briefly, we conduct 3
nocturnal broadcast surveys during the breeding period (April- August) across a network of
survey points to detect CSOs. When a CSO is detected we then conduct dusk status surveys
to pinpoint roost and nest locations for each bird. Status surveys are used to determine the
social status of each bird (pair or single), nesting and reproductive status (breeding, nonbreeding, unknown), and to identify color-banded individual birds.
In general, in years of higher CSO reproduction, such as occurred in 2004 and 2007, it is
easier to establish pair and reproductive status and to identify individual birds as they are
more vocal and exhibit stronger ties to their core areas. In years of lower reproduction, such
as occurred in 2005, 2006, 2008, 2009 (reproduction was low on the Plumas NF, while high
on the Lassen NF portion of the study area in 2009) and 2011, it is more difficult to
determine the status of birds as they tend to range more widely and are not as vocal and
territorial, particularly the females. Based on our cumulative survey results, we then use
accepted, standardized methods for estimating the overall number of territorial sites
(confirmed pairs, unconfirmed pairs and territorial singles) for each year. Confirmed pairs
consist of a reproductive pair of CSOs or, at non-reproductive sites, the detection of a male
and female on more than one occasion within 0.5 mile of each other across the breeding
period. Unconfirmed pairs consist of at least two detections of one sex but only a single
detection of the opposite sex within 0.5 mile of each other across the breeding period.
Territorial singles are individuals that are detected on at least 2 occasions within a 0.5 mile
distance across the breeding period without a detection of the opposite sex. Birds detected
on only a single occasion across the breeding period are not considered to be territorial.
Figure 8 illustrates treatment project areas and the cumulative number and distribution of
CSO territorial sites across the eight years between 2003 and 2011. The number of territorial
sites across SA-4 varied annually between 6 and 9 during 2003-2010 (Table 7, Fig. 9a&b).
Overall, the numbers of territorial sites were fairly similar with 7 sites documented between
2004-2006, an increase to 9 territorial sites during the high reproductive year that occurred in
80
2007, decreasing to 6 territorial sites in both 2008 and 2009, with 7 sites documented in
2010. However, in 2011 we documented 4 territorial sites within the MVPA (Table 7, Figure
9c.).
During 2008-2010 we did not detect much change in the number of CSO territorial sites
across SA-4, although we did document changes in occupancy status and spatial movements
of individual sites that may be associated with treatments. The Maple Flat site, located in the
NW corner of SA-4, was occupied from 2004 to 2007, not occupied in 2008 following
treatments in autumn 2007, and was colonized by an unconfirmed pair of new CSOs who
were present in the area during 2009 and 2010. Whether or not the treatments caused the site
to be unoccupied in 2008 is uncertain, as the male also died during the winter of 2007-2008
(determined by radio-telemetry). The female from 2007 visited the site in early 2008, then
moved and summered 14.5 km (9 miles) from Maple Flat near Seneca in early June 2008.
She remained in this area through October 2008 when she was recaptured and the radiotransmitter was removed. No new CSOs were detected or colonized the Maple Flat site in
2008.
We also observed movement of a site in the southwest corner of SA-4 that corresponded to
the timing of treatments in the nest core. This site (Miller Fork) was occupied by CSO pairs
in 2003-2005, a single male in 2006 following treatments in 2005-2006, and then was not
occupied between 2007 and 2009. However, a new CSO pair established a site (Big Creek)
in 2007 about 2 km to the northwest of this site. This new site has been occupied by a pair
from 2007 to 2010.
Changes in occupancy status were recorded at a third site that coincided with the timing of
treatments. A pair of CSOs occupied Whitlock Ravine from 2003 to 2008. Treatments were
implemented near the core area during 2008. The female CSO was found with a broken wing
alongside a road in November 2008. In 2009, the historic male occupied the site as a single
male. No CSOs occupied the site in 2010; the male was located 8.1 km (5 mi) southwest of
Whitlock Ravine.
CSO pairs were present at Deer Creek from 2003 to 2007 and fledged triplets in 2007. The
female died during the winter 2007-2008, and the site was not occupied from 2008 to 2010.
Treatments had been conducted to the south of the core area of this pair in 2006, and the
female was observed to forage in this area in 2007.
We also documented changes in occupancy at one site in the eastern portion of SA-4 that was
outside the area of treatments. Slate Creek was occupied by a male in 2003 and a pair in
2004, yet has not been occupied since. The female from 2004 was detected on the Lassen
NF, and then she was back on the Plumas in subsequent years. This female was not detected
in 2009 but was again observed in 2010.
We documented the colonization of a new site (Pineleaf Creek) during and following
treatments in the north portion of SA-4. A single territorial male was present in this area in
2006, with a pair of CSOs then present from 2007 to 2011. This pair successfully
reproduced in 2009. Although only a single example, it should be noted that this site
81
occurred in an area that had been underburned and the male of this pair (no radio on female)
preferentially foraged in this treatment type.
The Meadow Valley project area is the first area within the PLAS where full implementation
of HFQLG treatments has occurred. Annual monitoring of CSOs between 2003 and 2010
suggested that CSO numbers were fairly stable across the survey period, with the number of
pairs ranging between six and nine across years. Seven to nine territorial CSO sites were
documented each year between 2003 and 2007, whereas six territorial sites were located in
each of the first 2 years (2008 and 2009) following full implementation of the treatments
across the project area landscape. In 2010, seven territorial sites were documented in the
project area landscape. However, only four territorial sites were documented in 2011. Initial
post-treatment results through 2010 suggested that CSOs were able to persist in the
immediate post-treatment landscape, although movements of individual sites were
documented in areas that had experienced treatments, as well as in areas that had not
experienced treatments across the overall Meadow Valley project area. Radio-telemetry work
completed in 2010 indicated that CSOs avoid using recently treated Defensible Fuel Profile
Zone treatments. The decrease in number of CSO sites documented in 2011 may be
associated with a lag response of CSOs to treatment effects in the Meadow Valley Project
Area resulting from full treatment implementation. Alternatively, high winter snowfall may
have affected CSO numbers in the Meadow Valley Project Area. However, while high
snowfall was associated with near reproductive failure across the entire PLS study area, the
largest proportional decrease in number of territorial CSO sites was documented in the
Meadow Valley Project Area (Table 3). Information on the specific vegetation conditions
pre- and post-treatment across the study area, along with consideration of potential weather
interactions, is needed to more fully evaluate the association between CSO occupancy and
the effects of treatments to reach informed conclusions on treatment effects.
We recommend that monitoring be continued (1) to assess long-term occupancy, abundance,
and distribution of CSOs across the MVPA and PLS project area to document longer-term
responses to address concerns that site fidelity in such a long-lived species may obscure
possible negative effects of habitat change over the short term; and (2) to continue to monitor
color-banded birds to assess longer-term associations between CSO survival, reproduction,
and recruitment related to changes in habitat. Each of the pieces of above information is
necessary to fully assess the potential acute and chronic responses of CSOs to landscape
treatments. Accurate spatial post-treatment maps for the MVPA have been recently
completed (March 1, 2010) documenting: (1) the specific locations where treatments were
actually implemented on the ground; (2) the specific site-specific treatments that were
implemented on a piece of ground; and (3) when the treatments were implemented on the
ground (which year at minimum). Unfortunately, significant errors in the classification of
canopy cover within base vegetation maps were discovered in 2010 while conducting habitat
analyses and modeling exercises designed to explore the association between CSOs, their
habitat, and treatments. This resulted in the need to recreate an updated landscape vegetation
map for the project area. This updated vegetation map for MVPA was scheduled to be
completed and delivered by March 2011. To date the updated vegetation information is not
currently available. Understanding the what, where, when, and effects of treatments is the
foundation on which subsequent adaptive management assessments will be constructed. This
information has been difficult to obtain and is required to be able to explore the associations
82
between treatments and CSO responses at the landscape and home range spatial scales, in
addition to relating within home range habitat use through our telemetry studies.
Radio-Telemetry – Meadow Valley Project Area
Our landscape-scale research in the MVPA provides insight into CSO response at population
and territory spatial scales. Within home ranges, CSOs may also respond to treatments in
terms of selection of foraging habitat. To evaluate spotted owl foraging and home range
characteristics in a landscape recently modified by fuels-reduction treatments, we radiomarked and tracked 9 CSOs in the MVPA area from April 2007 to October 2008. Prior to
this study, spotted owl foraging patterns in post-treatment landscapes had not been welldescribed. This telemetry study was designed to explore initial behavioral responses of
California spotted owls to fuels treatments by characterizing home range configuration and
foraging site selection immediately following treatment installation.
We gathered 446 owl foraging locations across 2 breeding seasons and categorized forest
treatments into 4 types: Defensible Fuel Profile Zones (DFPZs), understory thin (removal of
trees <10-inches diameter); group selection (removal of all trees <30-inches diameter in
<0.8-ha patches); and understory thin followed by underburn. We estimated owl home
ranges from 30-60 owl locations per home range, using a fixed kernel density estimator; only
owl locations with an error ellipse <1.5 ha were used in analyses. We evaluated spotted owl
home range size and composition using repeated measures analysis of variance (ANOVA)
and resource selection functions. To analyze owl foraging patterns within home ranges, we
evaluated a priori hypotheses using an information-theoretic approach and Akaike’s
Information Criterion corrected for small sample sizes.
During the 2007 season, we gathered 236 nocturnal use locations across all individuals, with
4 locations occurring within fuels treatments. In 2008, we gathered 210 nocturnal use
locations, and 32 of these locations occurred within fuels treatments. Spotted owls used all
treatment types for nocturnal activities on at least one occasion; 51% of within-treatment
locations were accounted for by one spotted owl foraging repeatedly in underburn.
At the landscape scale, owl home ranges contained fuels treatments in proportion to their
availability on the landscape. Owl home ranges contained 7-35% fuels treatments (X=16%),
with home range size positively correlated with the total amount of fuels treatment within the
home range (p=0.049). Spotted owls selected against DFPZs (p=0.006), but not other fuels
treatments, for nocturnal activities; we hypothesize that the habitat character of DFPZs may
be unfavorable for common spotted owl prey species (Figure 11). One owl strongly selected
underburn treatments over untreated forest for foraging; limited availability of underburn
within the study area prevents further extrapolation of this result. Spotted owls foraged
much closer to their site center than expected by chance; because fuels treatments are not
permitted within PACs (located at most owl site centers), the required travel distance
between the site center and fuels treatments complicates result interpretation. Conclusions
from this study are exploratory and are intended to provide a baseline for further research
(Gallagher 2010).
83
Our next action, scheduled for completion in 2011, is to use a detailed landscape-scale
habitat map to incorporate habitat metrics into home range and foraging analyses. The plotscale vegetation structure and composition will also be analyzed at a subsample of CSO
radio-telemetry locations. Eighty-seven vegetation plots were measured to the standard FIA
protocol between August and November 2008, and 45 additional vegetation plots were
measured to the same protocol in September-October 2009.
We recommend further exploration of spotted owl use of fuels treatments, particularly
underburn, across multiple time periods and at patch, home range, and landscape spatial
scales. Additionally, considerations should be given to the design and implementation of
rigorous experimental studies to address the effects of fuels treatments on spotted owl
nesting and foraging habitat. A repeat of this study in 4-5 years, coupled with a study of
spotted owl population dynamics, would provide a comprehensive assessment of owl
response in the MVPA area. Evaluation of long-term effects is critical for long-lived species
such as the spotted owl (Blakesley et al. 2010); effects of fuels treatments on the owl may
manifest after short and long time periods, each with ramifications for ecological
understanding in the Sierra Nevada.
Wildfire – California Spotted Owl Case Studies
A primary source of uncertainty regarding the effects of fuels treatments is an assessment of
risk to CSOs and their habitat from treatments versus the risk from wildfire that occurs
across untreated landscapes. Prior to 2008 our PLS work had focused on assessing CSO
distribution, abundance and habitat associations across the untreated overall project area
landscape and being in position to monitor effects as treatments are implemented within
specific project areas, as illustrated by the MVPA case study described above. Beginning in
2008 we were fortunate to have the opportunity and funding support from the Plumas and
Lassen National Forests to extend our work to inventory CSO distribution, abundance, and
status across the Moonlight-Antelope Complex fire area (MACFA) that burned on the
Plumas National Forest in 2007. In 2009 we conducted a second year of surveys in the
MACFA and also conducted the first year of similar surveys in the Cub-Onion Complex fire
area (COCFA) that burned on the Lassen National Forest in 2008. As described below, the
MACFA was largely a high severity wildfire while the COCFA burned primarily at lowmoderate severity. In 2010 we conducted a second year of surveys in the COCFA.
Incorporating these two study areas which differ in wildfire severity allow us to directly
assess response of CSOs to landscapes that burned with different severities.
The MACFA consists of two fires that burned adjacent to each other in 2007, and both were
primarily high severity fires (Fig. 12). The MACFA covers approximately 88,000 acres. The
COCFA consist of two low-moderate severity fires that burned adjacent to each other during
June 2008 over approximately 21,000 acres (Fig. 12). About 52% of the MACFA burned at
high severity, whereas only 11% of the COCFA burned at high severity (Fig. 13).
In both wildfire study areas we conducted CSO surveys during the breeding period across the
entire landscape and within a 1.6 km (1 mile) unburned buffer surrounding the fire perimeter.
We used our standardized survey protocol and conducted three nocturnal surveys across the
landscape with follow-up visits to attempt to locate nest/roost locations for birds detected on
84
nocturnal surveys. These methods are described in the section above and fully in protocol
described in the study plan.
The high-severity fires that burned in the MACFA resulted in significant changes to the
vegetation (Fig. 14 & 15). The amount of suitable CSO habitat (CWHR classes 4M, 4D,
5M, 5D) within the 88,000 acre MACFA decreased from 70.1% of the pre-fire landscape to
5.8% of the landscape following the fires. The largest increase in the post-fire landscape
occurred in the CWHR classes <= 2D, which increased from 8.2% to 64.9%. The remaining
forested areas across the post-fire landscape were predominantly classified as either 4P
(18.5%) or 4S (7.9%).
We are still in the process of synthesizing all of the pre-fire CSO survey information for the
MACFA, as there is not a solid baseline of consistently collected survey information prior to
the fire such as exists for our core PLS project area. Nevertheless, this synthesis may
provide us with a reasonable estimate of the pre-fire distribution and abundance of CSO sites
across the MACFA. All or parts of at least 23 PACs were located within the pre-fire
MACFA. Given the lack of continuous annual CSO survey effort we are uncertain what
proportion of those PACs were occupied in 2007 prior to the fires.
During our 2008 surveys we documented a single confirmed pair of CSOs (non-breeding)
within the MACFA, with the female from this pair being the only female we detected within
the fire area (Fig. 16). We had 10 single detections of male CSOs across the burned area. In
each of these 10 cases we were not able to locate the birds at nests or roosts on follow-up
status surveys. Each of these 10 locations occurred primarily in the middle of the night when
birds are out foraging, and none of the detections occurred within 0.5 mile of each other as
required to classify these individuals as territorial birds under currently accepted protocols.
Within the unburned 1-mile buffer area surrounding the burned area we documented 5
confirmed pairs, 1 unconfirmed pair, 1 territorial male single, and 6 single detections (4
males, 2 sex unknown). Thus, in the immediate unburned buffer area we observed territorial
sites, whereas we only were able to document the single confirmed territorial pair within the
burned area.
During our 2009 surveys within the MACFA we documented a single confirmed pair of
CSOs in the same location as the pair documented in 2008 (Fig. 16). Within the 1-mile
buffer area we documented 7 confirmed pairs, 0 unconfirmed pairs, 2 territorial single males,
and 3 single detections. In contrast to 2008, in 2009 we did not record single detections of
apparently non-territorial single birds within the fire perimeter across the MACFA landscape.
Rather, we only recorded three detections of CSOs near the perimeter of the fire in the
vicinity of confirmed pairs located within the buffer around the fire.
In our two years of work we were able to document significant changes to the vegetation and
amounts and distribution of CSO habitat within the MACFA as a result of the high-severity
wildfires. Our CSO survey work suggests that the immediate post-fire landscape may not
support territorial CSO site,s as evidenced by the single confirmed pair of owls that we
documented in 2008. In 2009 we did not document any single male CSOs across the burned
landscape, suggesting that the apparently non-territorial single males observed in 2008 may
have been present because of previous site fidelity or were perhaps opportunistically utilizing
85
a flush of prey in the first year following the fire. In both years, territorial CSOs were present
in similar numbers and distributed at expected spacing within the buffer area surrounding the
fire. Thus, our results from our two years of work suggest that the primarily high-severity
MACFA does not support CSOs other than a single pair that is using the landscape. Further,
territorial CSO sites are well-distributed within the buffer area outside of the fire perimeter.
Our three detections of individual CSOs just within the perimeter of the burned areas suggest
that some CSOs are able to exploit the edge between the burned and unburned areas for
foraging.
In our first year of surveys during 2009 in the COCFA we documented 3 confirmed
territorial pairs, 1 unconfirmed territorial pair, and 2 territorial single male CSOs, for a total
of 6 territorial CSOs sites within the fire perimeter (Fig. 17). Additionally, we had 6 single
detections (3 male, 3 unknown sex) of individual CSOs within the fire perimeter. Within the
buffer area we documented 3 confirmed pairs and 3 single detections (2 male, 1 female).
These results and distribution patterns suggest that CSOs were able to persist in the post-fire
COCFA landscape with similar abundance and spacing as has been observed in unburned
forests outside the burned areas.
In 2010 we attempted to repeat the complete landscape survey coverage of the COCFA.
However, we were precluded from accessing the entire landscape area by law enforcement
restrictions due to safety concerns for field crews. Extensive, illegal marijuana growing
operations were distributed widely across the COCFA landscape. Due to this limitation we
instead conducted focused surveys within approximately 1-2 miles surrounding each of the
sites that were occupied by CSOs in 2009. Although we may have missed new sites
colonized in 2010 outside our focused survey areas, our efforts allowed us to determine if the
2009 sites were still occupied in 2010. In 2010 we documented 2 confirmed territorial pairs
and 1 unconfirmed territorial pair of CSOs within the fire perimeter, and 2 confirmed
territorial pairs and 1 unconfirmed territorial pair within the 1-mile buffer. Overall, in 2009
there were 7 pairs (6 confirmed and 1 unconfirmed) within the fire plus the 1-mile buffer. In
2010, there were 6 pairs (4 confirmed and 2 unconfirmed) in this same area. These results
suggest that CSOs are able to persist within landscapes that experience predominantly lowmoderate severity wildfire.
It is important to determine both the acute and chronic responses of CSOs and their habitat to
wildfire, as it is unknown if CSOs can persist over both the short term and long term in these
areas. Whether a post-fire landscape can support CSOs likely depends on the pre-fire habitat
suitability and variable fire severity patterns both within individual fires and across different
fires. Largely low-moderate severity fires may have positive or neutral effects on CSOs and
their habitat, while high severity fires may result in greater negative effects. Our results into
the acute, short-term response of CSOs to wildfire from the primarily high-severity MACFA
and primarily low-moderate severity COCFA support this hypothesis. Additionally,
information on the pre- and post-fire vegetation, location of salvage logging, and habitat
conditions are needed to fully assess the response of CSOs and their habitat to wildfire in the
MACFA and COCFA wildfire landscapes.
Storrie Fire Surveys
86
In 2011 we conducted a first year of CSO surveys within the Storrie Fire. Areas to be
surveyed were identified by Plumas and Lassen National Forest staff to facilitate
understanding fire effects on CSOs and to inform future restoration efforts. Not all of the
identified areas could be surveyed in 2011 due to lack of access resulting from a lingering
snowpack in the higher elevation, more remote regions of the targeted survey area. Surveys
were conducted across 76.7 km2, and a single territorial male CSO was documented (Fig.
18). Four additional single nocturnal CSO detections were documented; however, none of
the CSOs from these four detections could be located during follow-up surveys searching
for roosts and nests.
Banding, Blood Sampling, West Nile Virus Monitoring
Nineteen owls were captured and banded in 2011. Blood samples were collected from four
individuals that will be screened at the University of California, Davis for West Nile Virus
(WNV) antibodies. None of the 141 individual blood samples collected from the northern
Sierra Nevada from 2004 to 2008 have tested positive for WNV antibodies (Hull et al. 2010).
The 2009-2011 samples have not been analyzed to date.
Barred and Sparred (Spotted x Barred hybrid) Distributional Records
We detected 7 barred owl and 3 sparred owls during 2011 surveys within our PLS study area.
This is the same number of barred and sparred owls that we detected in 2010 and represents
the highest number of barred and sparred owls detected in any year during our study from
2004 through 2011. Our synthesis and update of barred-sparred owl records through 2011
based on Forest Service and California Department of Fish and Game databases indicates
that there are a minimum of 47 individual site records across the broader HFQLG Project
Area and a minimum total of 60 across the Sierra Nevada (Figure 19). This includes a
minimum total of 24 records that have been documented within our intensively surveyed
PLS study area. The first barred owl in the region was reported in 1989. The first
documented breeding in the PLS survey area was in 2007 and consisted of a sparred owl
paired with a CSO. In 2010, we documented the first barred-barred owl nest record for the
northern Sierra Nevada to the best of our knowledge. Despite several surveys, the outcome of
this barred-barred owl nesting attempt is uncertain, as the pair was non-responsive during the
fledgling period. The pattern of records suggests that barred/sparred owls have been
increasing in the northern Sierra Nevada from 1989 to 2011 and are now present in low,
stable numbers over the past 4-5 years on our study area. This pattern is consistent with that
observed in other areas as barred owls have expanded their range in western North America.
Initially barred owls colonize an area and persist at low population numbers; during this
period they may hybridize with spotted owls. At some threshold population size or when
ecological conditions allow they are then poised for, and capable of, exhibiting exponential
population growth. Notably, we detected an increase in barred numbers in 2010 and 2011.
This situation requires further monitoring, and we recommend that an inventory of barred
owls be conducted using barred owl-specific survey methods to document the distribution
and status of barred owls in the northern Sierra Nevada.
California Spotted Owl Diet
87
A single diet survey plot was established at a CSO nest or roost location at each CSO
territory on the Plumas National Forest during 2003-2007. Systematic searches for pellets
and prey remains were conducted in each plot during each year. A total of approximately
3398 pellets have been collected during 2003-2007 (2003 = 606; 2004 = 807; 2005 = 838;
2006 = 516; 2007 = 552). We completed sorting of all pellets and identification of all prey
remains in January 2010. All prey items were identified to species, or to taxonomic group
when species identification could not be ascertained. A total of 8,595 prey items have been
recorded from the pellets. Mammals are the dominant taxonomic group and comprise 96.5%
of the total biomass identified in the diet. Across years the highest biomass contributions
were from the dusky-footed woodrat (contributed 45% of the estimated total biomass) and
northern flying squirrel (10.8%). Our objective has been to sample over several years to
assess temporal variation in diets and possible relationships to variation in CSO
reproduction, and to sample widely over space in order to investigate potential variation in
CSO diets associated with elevation and vegetation conditions. We will now be able to
address these questions given completion of the sorting and prey identification from the
pellet samples.
Summary 2003-2011
Our efforts from 2003 to 2011 have focused on monitoring CSO distribution, abundance and
demographics to address our primary research objectives and provide the baseline data for
assessing the effects of HFQLG implementation. In conjunction with the now fully
integrated Lassen Demographic Study, we have collected landscape-scale information on the
distribution and abundance of CSOs across approximately 650,000 acres of land.
Determining the accurate number and distribution of CSO sites requires multiple years of
survey and marking of individual CSOs to delineate separate territories and identify
individual birds that move among multiple sites within and across years. These baseline data
are fundamental for developing empirically-based habitat models for understanding CSO
habitat associations and developing adaptive management tools and models. The completion
of the Meadow Valley area projects in 2007-2008 marked the first landscape series of
HFQLG treatments to be implemented within the study area, providing the first opportunity
to address treatment effects within a case study framework. Our baseline information on
CSO distribution and habitat associations, coupled with our 2007-2008 radio-telemetry work,
will allow us to assess associations between CSOs and vegetation changes. In 2008-2011 we
were now able to monitor post-treatment CSO distribution and abundance in the Meadow
Valley project area, providing the first empirical data from a treated landscape. Our shortterm results to date suggest that CSOs are able to persist in this treated landscape, although
we have observed territory-specific responses in both areas that have experienced treatments
and sites where no treatments have occurred. In 2010, we also completed the analysis of the
radio-telemetry work, which has provided insight into how individual owls respond to
treatments within the first 1-2 years following implementation.
Additionally we were able to expand our work to address the effects of wildfire on CSOs and
their habitat through our 2008-2010 survey work in the Moonlight-Antelope Complex and
Cub-Onion Complex fire areas. In summary, we are working towards being able to broadly
address CSO management questions across a gradient of landscape conditions ranging across
88
untreated landscapes, landscapes treated to meet desired fuels/vegetation conditions, and
landscapes that have experienced wildfire in order to address primary management issues.
Dedicated monitoring of CSOs on the Lassen Demographic Study continues to provide
critically valuable demographic and population trend information for determining the status
of CSOs. Our analyses of the demographic data through 2011 continue to suggest that the
CSO population on study area declined over the 20-year study period, similar to the results
through 2005 reported in Blakesley et al. (2010). These results warrant close continued
monitoring of the status of CSOs within the study area, along with continued management
focus on providing high-quality CSO habitat during the planning and implementation of
HFQLG treatments. We are just now reaching a study duration where we can begin to
estimate population trends for the Plumas NF study areas. To date, our baseline information
on CSO distribution and abundance suggests that numbers of territorial CSOs and sites have
been fairly consistent from 2004-2011, with a drop in numbers in 2011.
Our focused diet analyses have broadened and deepened our understanding of CSO diets and
sources of variation in CSO diets among pairs and across environmental gradients.
Monitoring of WNV exposure coupled with demographic monitoring has provided an
opportunity to assess if WNV may ultimately be a factor influencing CSO viability. We have
provided the first information investigating evidence for CSO exposure to West Nile Virus in
the Sierra Nevada (Hull et al. 2010). To date we have not had a positive detection for WNV
within CSOs. Through our research into historical and current occurrence records, in
conjunction with our field surveys, we have been able to document the colonization of the
northern Sierra Nevada by barred owls. Our results indicate that barred owls are increasing in
the northern Sierra Nevada and may become an increasing risk factor for CSOs.
Current Research: 2012
In 2012 we will continue monitoring owl distribution, abundance, demography, and
population trend across the Lassen Demographic Study Area and in the Meadow Valley
Project Area. This monitoring includes the Creeks Project Area on the Lassen NF, as this
area is projected to receive forest and fuels treatments in the near future. Additionally, we
will attempt, pending time and resources, to monitor CSOs across the remainder of the core
PLS Area in an attempt to provide full consistent information across the entire study area
that will be available for habitat modeling and continuation of baseline monitoring. In 2012
we will also conduct a second year of surveys in suitable CSO habitat within the Storrie Fire
footprint to collect information on CSO response to this wildfire following 10-11 years of
post-fire vegetation change. These data will also be used to inform restoration efforts within
the Storrie Fire landscape. Together this work will contribute to our efforts to build a more
comprehensive base of knowledge regarding CSO habitat associations and the effects of
treatments and wildfires.
In addition to continuing field surveys designed to address our seven research questions, we
have broadened our emphasis on the development of predictive habitat relationship models
as described in the module study plan. We have continued to work closely with biologists on
the Plumas and Lassen National Forests, and the R5 Regional Office, to identify and define
the types of analyses and tools that would best address management needs. Baseline
89
information collected during this study forms the foundation for this phase of the research.
The combination of broad-scale landscape CSO distribution data, in conjunction with
detailed demographic information available from the Lassen Demographic Study, will
facilitate exploration and development of predictive habitat models for use in an adaptive
management framework and to directly monitor implementation of the HFQLG project.
There remain two challenges to completing and leveraging this research to fully realize the
potential of this work to advance our knowledge on CSO associations with vegetation,
treatments and wildfire. First, the lack of available, and delays in obtaining, updated
vegetation maps of accurate pre- and post-treatment vegetation information and accurate
spatial locations of treatments remains a significant roadblock. Usable vegetation maps have
not been completed nor been available for our efforts to address vegetation and treatment
mapping information needs for habitat modeling. Completion of the needed vegetation maps
will facilitate completion of the primary habitat modeling research objectives and to assess
treatment and wildfire effects on CSOs and their habitat for this project. Detailed information
on the post-treatment vegetation responses is needed to more fully explore the effects of
treatments on CSOs within the Meadow Valley Project Area and across the broader PlumasLassen Study Area. Second, continued monitoring of the Meadow Valley Project Area is
required to assess if the decrease in CSO sites observed in 2011 persists over time. Further,
treatment and monitoring of additional treated landscapes is needed to provide replicate case
studies to determine if similar results are broadly documented in other situations. The strong
baseline data on CSO distribution established across the PLS study area provides a unique
and valuable opportunity to implement additional landscape forest restoration and fuels
treatments to assess effects on CSOs and their habitat.
Literature Cited:
Blakesley, J.A., M.E. Seamans, M.M. Connor, A.B. Franklin, G.C. White, R.J. Gutierrez,
J.E. Hines, J.D. Nichols, T.E. Munton, D.W.H. Shaw, J.J. Keane, G.N. Steger, B.R. Noon,
T.L. McDonald. 2010. Population dynamics of spotted owls in the Sierra Nevada, California.
Wildlife Monographs No. 174.
Gallagher, C.V. 2010. Spotted owl home range and foraging patterns following fuelsreduction treatments in the northern Sierra Nevada, California. MSc. Thesis. University of
California, Davis. 54p.
Hull, J.M., J.J. Keane, L.A. Tell, and H.B. Ernest. 2010. West Nile virus anti-body
surveillance in three Sierra Nevada raptors of conservation concern. Condor 112(1):168172. (Refereed).
90
91
Table 1. California spotted owl reproduction on the Plumas and Lassen National Forests
2004-2010.
Year
2004
2005
2006
2007
2008
2009
2010
2011
Percent of confirmed/unconfirmed pairs
with successful nests
49.4%
17.7%
13.8%
55.4%
16.4%
47.6%
50.8%
2.0%
Young fledged per
successful nest
1.68
1.47
1.50
1.81
1.70
1.57
1.74
2.00
92
Table 2. Crude density of territorial California spotted owls across survey areas on the Plumas and Lassen National Forests 2004-2011.
Locations of survey areas are identified in Figure 1.
Crude Density of Territorial Owls (#/km2)
Survey Area
SA-2
SA-3
SA-4
SA-5
SA-7
SA-1Ab
SA-1Bb
SA-11b
SA-12b
SA-13b
SA-14b
SA-15b
Total Study
Area
Size
(km2)
182.4
214.4
238.2
260.2
210.3
190.4
130.3
179.4
215.8
152.9
318.7
196.8
2004a
2005a
2006a
2007a
2008a
2009a
2010a
2011a
0.126
0.075
0.059
0.069
0.071
NI
NI
NI
NI
NI
NI
NI
0.143
0.093
0.050
0.069
0.062
0.042
0.023
0.045
0.097
0.105
0.053
0.086
0.115
0.089
0.046
NS
NS
0.042
NS
0.033
0.070
0.085
0.044
0.036
0.115
0.103
0.071
NS
NS
0.053
NS
0.033
0.074
0.065
0.035
0.056
0.132
0.098
0.046
NS
NS
0.042
NS
0.045
0.070
0.050c
0.047
0.081
0.121
0.089
0.046
NS
NS
0.058
NS
0.033
0.074
0.099
0.044
0.076
0.132
0.079
0.050
NS
NS
0.053
NS
0.039
0.102
0.099
0.035
0.076
0.104
0.070
0.029
NS
NS
0.032
NS
0.039
0.074
0.078
0.035
0.097
2489.8
0.078
0.073
0.060
0.066
0.067
0.069
0.070
0.059
NI = Not included. Project level area surveyed only in 2005; included for comparative purposes.
NS= Not surveyed.
a
Total Area surveyed each year: 2004 = 1,106 km2; 2005 = 2,490 km2; 2006 = 1,889 km2; 2007 = 1,889 km2; 2008 = 1,877 km2;
2009-2011 = 1,889 km2
b
The Lassen Demographic Study Area (SA-1A, SA-1B, SA-11 through SA-15) was incorporated into the overall study in 2005.
c
This survey area was not completely surveyed during 2008 because of wildfire activity in the area. Two CSO territories within the
study area could not be surveyed.
93
Table 3. Number of pairs (confirmed and unconfirmed) and territorial single California spotted owls across the Plumas-Lassen Study
survey areas on the Plumas and Lassen National Forests, California, 2004-2011. Locations of survey areas are identified in Figure 1.
Survey
Area
SA-2
SA-3
SA-4
SA-5
SA-7
SA-1Aa
SA-1Ba
SA-11a
SA-12a
SA-13a
SA-14a
SA-15a
2004
Pairs/ TS
2005
Pairs/TS
2006
Pairs/TS
2007
Pairs/TS
11/1
7/2
7/0
8/2
7/1
NI
NI
NI
NI
NI
NI
NI
12/2
10/0
5/2
9/0
6/1
4/0
3/0
4/0
10/1
8/0
8/1
8/1
10/1
9/1
4/3
NS
NS
4/0
NS
3/0
1/7
6/1
7/0
3/1
10/1
11/0
8/1
NS
NS
5/0
NS
3/0
8/0
5/0
5/1
4/3
2008
Pairs/TS
12/0
9/3
5/1
NS
NS
4/0
NS
3/2
7/1
3/1b
7/1
8/0
2009
Pairs/TS
2010
Pairs/TS
2011
Pairs/TS
11/0
9/1
5/1
NS
NS
5/1
NS
3/0
8/0
7/0
7/0
7/1
10/2
8/1
5/2
NS
NS
4/2
NS
3/1
10/2
7/0
5/1
7/1
9/1
7/1
3/1
NS
NS
1/4
NS
3/1
6/4
5/2
5/1
9/1
NI = Not included. Project level area surveyed only in 2005; included for comparative purposes.
NS= Not surveyed.
TS = Territorial Single.
a
The Lassen Demographic Study Area (SA-1A, SA-1B, SA-11 through SA-15) was incorporated into the overall study in 2005.
This survey area was not completely surveyed during 2008 because of wildfire activity in the area. Two CSO territories within the
study area could not be surveyed.
b
94
Table 4. Distribution of California spotted owl nest/primary roost sites (n = 103) across
CWHR tree size classes within the Plumas-Lassen Study on the Plumas and Lassen National
Forests, 2004-2006.
CWHR
CWHR Size Class Description
Number Percent
Size
of Nests
Class*
Barren
Open, sparse tree coverage
1
1.0
3S
6-12 inch dbh, ,20% CC
1
1.0
3M-LT 6-12 inch dbh, 40-60% CC, large trees recorded
1
1.0
3D
6-12 inch dbh, >60% CC
4
3.9
4P
12-24 inch dbh, 20-40% CC
3
2.9
4M
12-24 inch dbh, 40-60% CC
3
2.9
4M-LT 12-24 inch dbh, 40-60% CC, large trees recorded
12
11.7
4D
12-24 inch dbh, >60% CC
10
9.7
4D-LT 12-24 inch dbh, >60% CC, large trees recorded
13
12.6
5M
>24 inch dbh, 40-60% CC
25
24.3
5D
>24 inch dbh, >60% CC
9
8.7
6
>24 inch dbh, >60% CC, multi-layer canopy
21
20.1
*defined by average tree size (dbh = diameter at breast-height) and average percent
canopy cover (CC).
95
Table 5. Nest-site (1 ha (2.47 acres)) habitat characteristics collected using the Forest
Inventory and Analysis sampling protocol at California spotted owl nest sites (n = 80) on the
Plumas and Lassen National Forests, California, 2005-2006.
Variable
Mean
SE
Total Basal Area (ft2/acre)
260.8
6.47
# Trees >= 30 inch dbh (#/acre)
10.7
0.58
2
Basal Area Trees >= 30 inch dbh (ft /acre)
96.0
5.70
# Trees >= 24 inch dbh (#/acre)
19.9
0.90
2
Basal Area Trees >= 24 inch dbh (ft /acre)
131.7
6.29
# Trees <12 inch dbh (#/acre)
383.5
26.36
2
Basal Area Trees , <12 inch dbh (ft /acre)
50.1
2.71
# Snags >=15 inch dbh (#/acre)
7.4
0.80
Mean Duff Depth (inches)
3.0
0.16
Duff (tons/acre)
67.4
3.64
Mean Litter Depth (inches)
2.3
0.18
Litter (tons/acre)
23.7
1.81
1 Hour Fuels (tons/acre)
0.75
0.03
10 Hour Fuels (tons/acre)
4.0
0.21
100 Hour Fuels (tons/acre)
4.4
0.28
Shrub Cover (%)
7.7
1.16
Canopy Cover (%)*
64.1
1.24
* estimated through Forest Vegetation Simulator modeling of plot-based tree lists.
96
Table 6. Distribution of USDA Region 5 vegetation classes (Mean (SE)) within 500 acre
(201 ha) circles centered on California spotted owl (CSO) territories (n = 102) and
systematic grid (Grid) points (n = 130) within the Plumas-Lassen Study on the Plumas and
Lassen National Forests, 2004-2006.
R5 Size
Class*
Non-forest
Total Size 1
R5 Size Class Description
CSO
Grid
Sum of non-forest land types
Sum of 1G,1N, 1P, 1S: <6 inch dbh,
all %CC classes
6-12 inch dbh, 10-39% CC
6-12 inch dbh, 40-69% CC
6-12-24 inch dbh, >=70% CC
12-24 inch dbh, >10-39% CC
12-24 inch dbh, 40-69% CC
12-24 inch dbh, >=70% CC
>24 inch dbh, >10-39% CC
>24 inch dbh, 40-69% CC
>24 inch dbh, >=70% CC
Sum of 4N & 4G: >24 inch dbh, >=
40% CC
Sum of classes 3N, 3G, 4N, 4G =
>12 inch dbh, >40% CC
4.4 (1.0)
1.7 (0.3)
8.4 (1.2)
1.6 (0.3)
2P & 2S
3.4 (0.4)
4.1 (0.5)
2N
3.8 (0.6)
4.4 (0.9)
2G
1.6 (0.5)
0.5 (0.1)
3P&3S
9.2 (0.8)
16.1 (1.3)
3N
37.2 (2.4)
38.5 (1.8)
3G
6.2 (1.0)
3.8 (0.7)
4P&4S
1.0 (0.3)
2.1 (0.4)
4N
25.8 (2.0)
17.3 (1.6)
4G
6.5 (0.1)
2.4 (0.8)
Total 4N &
32.4 (2.3)
19.6 (1.8)
4G
Total
Suitable
75.7 (2.19)
61.9 (1.75)
habitat
*defined by average tree size (dbh = diameter at breast-height) and average percent
canopy cover (CC).
97
Table 7. Annual number of California spotted owls documented during the breeding period
(April-August) in SA-4 (Meadow Valley Project Area), Plumas National Forest, California,
2003-2010.
Year
2003
2004
2005
2006
2007
2008
2009
2010
2011
Confirmed
Pairs
7
7
4
3
8
5
5
5
3
Unconfirmed
Pairs
0
0
1
1
0
0
0
0
0
Territorial
Singles
1
0
2
3
1
1
1
2
1
Total Territorial
Sites
8
7
7
7
9
6
6
7
4
Figure 1. (A) Location of California spotted owl (CSO) Survey Areas surveyed
2004-2011. (B) Example of original survey plot consisting of multiple Cal-Planning
watersheds. (C) Example of Primary Sampling Units for surveying for CSOs. See
text and study plan for further details.
Figure 2. Distribution of California spotted owl territories within the study survey area
across the Plumas and Lassen National Forests, 2011.
Figure 3. Monthly precipitation totals for Quincy, California, during January- May, 2005-2011
(data from Western Regional Climate Center).
Figure 4. Estimates of realized population change (realized lambda) for the Lassen
Demographic Study Area, 1990-2011.ab
a
The dashed line at 1.0 represents a stable population that is neither increasing nor declining.
The first two and last year estimates are removed because they are biased due to the statistical
methods used.
b
Figure 5. Distribution of California spotted owl (n = 103) nest sites by California Wildlife
Habitat Relationship (CWHR) database vegetation classes on the Plumas and Lassen National
Forests, California, 2004-2007. Descriptions of the CWHR classes are provided in Table 5
within the text of this document.
30
25
# nests
20
15
10
5
0
BAR
3S
3MLT
3D
4P
4M
4MLT
4D
4DLT
Size/Density type of nest's veg polygon
5M
5D
6
Figure 6. Percent suitable habitat (>=12 inch dbh trees with >=40% canopy cover) within 500 acre (201
ha) circles centered on California spotted owl (CSO, n = 102) and systematic grid points (Grid, n = 130)
on the Plumas and Lassen National Forests, California, 2004-2007.
35
30
Percent of Observations
25
20
CSO
GRID
15
10
5
0
0-20
21-30
31-40
41-50
51-60
61-70
71-80
Percent Suitable Habitat (3N, 3G, 4N, 4G)
81-90
91-100
Figure 7. Percent large tree habitat (R5 classes 4N &4G: >=24 inch dbh trees with >=40%
canopy cover) within 500 acre (201 ha) circles centered on California spotted owl (CSO, n =
102) and systematic grid points (Grid, n = 130) on the Plumas and Lassen National Forests,
California, 2004-2007. Descriptions of R5 classes are provided in Table 7 within the text of
this document.
20
15
10
5
0
0%
1-10%
11-20%
21-30%
31-40%
41-50%
51-60%
61-70%
71-80%.
81-90%
Figure 8. Distribution of proposed Meadow Valley Project Area forest management
treatments and cumulative distribution of California spotted owl territorial sites between
2003 and 2010 in Survey Area-4 of the Plumas-Lassen Study, Plumas National Forest,
California.
Figure 9a. Annual summary distribution of California spotted owl territorial sites across
Survey Area-4 (Meadow Valley Project Area) of the Plumas-Lassen Study, Plumas National
Forest, California, 2003-2006.
Figure 9b. Annual summary distribution of California spotted owl territorial sites across
Survey Area-4 (Meadow Valley Project Area) of the Plumas-Lassen Study, Plumas National
Forest, California, 2007-2010.
Figure 9c. Annual summary distribution of California spotted owl territorial sites across
Survey Area-4 (Meadow Valley Project Area) of the Plumas-Lassen Study, Plumas National
Forest, California, 2011.
Figure 10. Number of Spotted Owl territories in the Meadow Valley Project area versus the
overall number of territories in the Plumas-Lassen Study area.
a
Gray line at 2006 represents the year in which implementation of large-scale fuels treatments began on the
ground.
Figure 11. Average California spotted owl use locations compared with random for four forest
treatment types in the Meadow Valley Project area, Plumas National Forest, California, 20072008a.
30
25
20
Owl Use
15
Ra ndom
10
5
0
DFPZ
a
Understory thin
Group selection
Underburn
Individual owls were considered the sample unit and not all treatment types were available to each
owl: Defensible Fuel Profile Zone (DFPZ; n = 9); understory thin (n = 5); group selection (n = 9);
and understory thin followed by underburn (n = 4).
Figure 12. Maps of fire severity in the: (a) Moonlight-Antelope Complex fire (88,000 acres) of 2007; and (b) the Cub- Onion
Complex fire (21,000 acres) of 2008, on the Plumas and Lassen National Forests, California.
(a)
(b)
111
Figure 13. Distribution of the post-fire landscape by fire severity class for the MoonlightAntelope Complex Fire Area (MACFA) and Cub-Onion Complex Fire Area (COCFA) on the
Plumas and Lassen National Forests, California.
70
60
Percent
50
40
30
MACFA
COCFA
20
10
0
High
Moderate
Low
Fire Severity
Unburned
Figure 14. Distribution of pre- and post-fire California Wildlife Habitat Relationship vegetation
classes within the Moonlight-Antelope Complex fire areas on the Plumas and Lassen National
Forests, California, 2008.
70
60
Percent
50
40
Pre-Fire (%)
30
Post-Fire (%)
20
10
0
<=
2D
3D
3M
3P
3S
4D
4M
4P
CWHR Class
4S
5D
5M
5P
5S
Figure 15. Maps of (a) pre-fire and (b) post-fire California Wildlife Habitat Relationship
vegetation classes within the Moonlight-Antelope Complex fire areas on the Plumas and Lassen
National Forests, California, 2008.
(a)
(b)
Figure 16. Distribution of California spotted owls within the Moonlight-Antelope Complex
fire area and a 1.6 km buffer on the Plumas and Lassen National Forests, California, 20082009.
Figure17.
15.Distribution
DistributionofofCalifornia
Californiaspotted
spottedowls
owlsdetected
detectedinin(a) 2009 and (b) 2010 within four wildfire burn severity classes in
Figure
(a)Cub-Onion
2009 and (b)
2010 within
the Cub-Onion
Complex
the
Complex
Fire Area
with a 1.6 km
buffer,Fire
Lassen National Forest, California.
Area and a 1.6 km buffer and wildfire burn severity classes on
the Lassen National Forest, California.
(a)
(b)
117
Figure 18. Results of CSO surveys within the Storrie Fire (2000) during 2011. High priority
survey areas were identified by Plumas and Lassen National Forest personnel due to extreme
difficulty of the terrain.
118
Figure 19. Distribution of Barred and Sparred (Spottex-Barred Hybrid) Owls within the HFQLG
Project area, 1989-2011.
119