# Endangered and Threatened Wildlife and Plants; 12-Month Finding for a Petition To List the California Spotted Owl (Strix occidentalis occidentalis)

> Briefs, arguments, decisions, and more.

URL: https://www.frixlaw.com/law-library/documents/fr%3A03-3519

## Record

- **Collection:** Federal Register
- **Document type:** Notice
- **Published:** February 14, 2003
- **Citation:** 68 FR 7580

## Text

DEPARTMENT OF THE INTERIOR
Fish and Wildlife Service

Endangered and Threatened Wildlife and Plants; 12-Month Finding for a Petition To List the California Spotted Owl (
Strix occidentalis occidentalis
)

AGENCY:

Fish and Wildlife Service, Interior.

ACTION:

Notice of 12-month petition finding.

SUMMARY:

We, the U.S. Fish and Wildlife Service (Service), announce a 12-month finding for a petition to list the California spotted owl (
Strix occidentalis occidentalis
) under the Endangered Species Act of 1973, as amended. After reviewing the best available scientific and commercial information available, we find that the petitioned action is not warranted. We continue to ask the public to submit to us any new information that becomes available concerning the status of or threats to this species. This information will help us monitor and encourage the conservation of this species.

DATES:

The finding announced in this document was made on February 7, 2003. You may submit new information concerning these species for our consideration at any time.

ADDRESSES:

You may send data, information, comments, or questions concerning this finding to Field Supervisor (Attn: CASPO), Sacramento Fish and Wildlife Office, U.S. Fish and Wildlife Service, 2800 Cottage Way, Room W-2605, Sacramento, California 95825. You may inspect the petition, administrative finding, supporting information, and comments received, by appointment, during normal business hours, at the above address.

FOR FURTHER INFORMATION CONTACT:

Susan Moore or Ken Sanchez at the above address (telephone at 916/414-6600; facsimile at 916/414-6710).

SUPPLEMENTARY INFORMATION:

Background

Section 4(b)(3)(B) of the Endangered Species Act of 1973, as amended (Act) (16 U.S.C. 1531
et seq.
), requires that, for any petition to revise the List of Threatened and Endangered Species that presented substantial scientific or commercial information that listing may be warranted, we make a finding within 12 months of receiving the petition on whether the petitioned action is: (a) Not warranted, (b) warranted, or (c) warranted but precluded by other pending proposals. Such 12-month findings are to be published promptly in the
Federal Register
.

On April 3, 2000, we received a petition dated April 2000, from the Center for Biological Diversity, Tucson, Arizona, and Sierra Nevada Forest Protection Campaign, Sacramento, California, and other organizations to list as threatened or endangered the California spotted owl (
Strix occidentalis occidentalis
). The names, addresses, and signatures of representatives of these organizations followed in a letter dated April 17, 2000. These organizations filed the petition on behalf of themselves and 14 other organizations and requested that we designate critical habitat for the California spotted owl concurrent with listing. Further, they requested emergency listing and emergency designation of critical habitat. Although emergency listing and designation of critical habitat are not petitionable actions under the Act, we determined that an emergency situation did not exist.

On October 12, 2000, we published a 90-day finding on that petition in the
Federal Register
(65 FR 60605). In that publication we found that the petition

presented substantial scientific or commercial information to indicate that listing the California spotted owl may be warranted, and we requested information and data regarding the species. On July 31, 2001, the Center for Biological Diversity and others filed a complaint in District Court, alleging the Service failed to make a timely 12-month finding in response to their listing petition. On March 5, 2002, the District Court entered an order requiring the completion of the 12-month finding by February 10, 2003.

The Petition

The petitioners believe that listing the California spotted owl is necessary because of factors related to loss and modification of habitats from timber harvest and urbanization, lack of existing State or Federal regulatory mechanisms that protect the species, and declines in the population.

The petitioners believe listing is necessary primarily because past timber harvest in the Sierra Nevada has resulted in the loss of key components of spotted owl habitat over large portions of the landscape. They also believe that current Federal land management agency strategies and private land forest practices are resulting in the loss or destruction of spotted owl habitat. They expressed special concern about past timber harvest practices that selectively removed the larger, older trees that comprise a key component of spotted owl breeding habitat.

The petitioners refer to the “Interim Guidelines” of the Federal land management strategy in place at the time the petition was submitted (April 2000) to conclude that current and planned timber sales would continue to remove key components of spotted owl habitat. The petitioners believe the cumulative effects of continued timber harvest and fuels reduction projects on Federal lands would have dramatic effects on the spotted owl.

The petitioners state, “* * * there are almost no protections for spotted owls * * *” on private lands. They assessed the State mechanism for permitting timber harvest and analyzed recent timber harvest plans to conclude, “* * * owls are being heavily impacted by logging on private lands.” Additional evidence of habitat destruction cited by the petitioners includes urbanization and development, particularly loss of habitat at lower elevations from new home construction.

The petitioners cite recent studies that report potential population declines as further evidence to support a positive listing decision. The petitioners review and interpret several studies of California spotted owl population dynamics to infer “drastic” annual declines in the population.

Other impacts addressed in the petition include livestock grazing, recreation, climate change, fire, competition from the barred owl, and disease and predation. These are impacts thought by the petitioners to be apparent, though not well studied or documented.

Taxonomy and Description

The spotted owl was first described as
Syrnium occidentale
by John Xantus in 1859 based on a specimen collected at Fort Tejon, Kern County, California (Xantus 1859). The species was later reassigned to the genus
Strix
(Ridgway 1914). The specific name was altered to conform to the Code of Zoological Nomenclature, yielding the scientific name
Strix occidentalis
(Service 1993). Currently, the American Ornithologist Union (AOU) recognizes three subspecies of spotted owls: the California spotted owl (
Strix occidentalis occidentalis
), the northern spotted owl (
Strix occidentalis caurina
), and the Mexican spotted owl (
Strix occidentalis lucida
) (AOU 1957).

The spotted owl is mottled in appearance. It has a brown back with white spots and brown barring. The facial disk is pale brown with concentric rings of dark brown, bordered by a ring of dark brown feathers. A conspicuous light-colored “X” is apparent between the eyes above its pale yellowish beak, where “eyebrows” and “whiskers” merge together. Unlike most other owl species, which have yellow eyes, spotted owls have dark brown eyes. Wings and tail are rounded, and all flight feathers are dark brown with light brown cross-bars. Sexes cannot be distinguished by plumage, but can be readily identified by size and vocalization (Verner
et al.
1992b). Females are usually larger than males, with males weighing 470 to 685 grams (g) (17 to 24 ounces (oz)), and females 535 to 775 g (19 to 27 oz) (Gutie

rrez
et al.
1995). First- and second-year adults can be distinguished by the tips of the tail feathers, which are white and taper to a sharp point until replaced by adult plumage at about 26 months of age (Gutie

rrez
et al.
1995). The spotted owl is the fifth largest species of owl occurring in North America (Verner
et al.
1992b); It is 41 to 48 centimeters (cm) (16 to 19 inches (in)) in length, with a wingspan of 107 to 114 cm (42 to 45 in) (Center for Biological Diversity 2000).

California spotted owls are lighter brown with slightly larger white spots than the northern spotted owl. Mexican spotted owls are lighter brown than both the California and northern subspecies, with some individuals having a rare palomino color. The facial disk and upper breast of the Mexican spotted owl contain more white than the other subspecies, and larger white spots add to the perception that they are lighter in color (Gutie

rrez
et al.
1995).

The northern spotted owl was listed as threatened under the Act in 1990 (Service 1990), and the Mexican spotted owl was listed as threatened in 1993 (Service 1993).

Population Genetics

Three genetic markers (
i.e.
, allozymes, mitochondrial DNA and random amplified polymorphic DNA) have been used to examine the genetic structure of spotted owls . Analysis of allozymes (alternate forms of proteins) supports separation of the Mexican spotted owl from the other two subspecies (Gutiérrez
et al.
1995). Barrowclough
et al.
(1999) compared the sequences of a fragment of mitochondrial DNA (mtDNA) from 73 individual spotted owls, including samples from all three subspecies and from multiple populations within each subspecies. Their data support the separation of the species into the three currently recognized subspecies. Based on their data, the northern spotted owl appears to have diverged from the other two subspecies, and the California spotted owl later diverged from the Mexican spotted owl. In this study, gene flow appeared relatively high within subspecies and low between subspecies (Barrowclough
et al.
1999). The authors concluded that gene flow between northern and California spotted owls is a recent and uncommon phenomenon.

Haig
et al.
(2001) used random amplified polymorphic DNA (RAPD) to analyze genetic variation between spotted owls at multiple geographic levels, including between subspecies. They found extremely low RAPD variation in spotted owls, with only 11 of 400 primers showing variation. Their data show genetic separation of Mexican spotted owls from California and northern spotted owls, but do not show separation between the California and northern subspecies. They suggest that the lack of separation between the California and northern subspecies in their data may be due to recent gene flow between subspecies, or due to the low variation of the data. We are also aware that additional research by Haig and colleagues bearing on the question of subspecific distinctions in spotted owls has not yet been published (A. Bowers, U.S. Fish and Wildlife Service, in litt. 2002).

Currently available, published genetic data (
i.e.
, mtDNA and RAPDs) apparently lead to different conclusions regarding subspecific distinctions in spotted owls. Therefore, for the purposes of this finding, we adopt the taxonomy accepted by the American Ornithological Union (AOU 1957), which recognizes the California spotted owl as a distinct subspecies (
Strix occidentalis occidentalis
).

Life History

Mating System and Reproduction.
Spotted owls usually reach reproductive maturity at two years of age, although first year birds have sometimes nested the season after they were hatched. Considerable variation exists in both the percentage of pairs that nest and the number of pairs that successfully fledge young, both geographically and from year to year (Verner
et al.
1992b).

Spotted owls are monogamous with no records of extra-pair copulations. They usually pair with the same mate from year to year, although “divorces” have been documented. The breeding season of California spotted owls extends from mid-February to mid-September or early October. Individuals begin breeding earlier in the San Bernardino Mountains than in the Sierra Nevada. Within a geographic area, individuals begin breeding earlier at lower elevations (Verner
et al.
1992b).

California spotted owls are mostly nonmigratory, remaining within the same home ranges year round. However, in the Sierra Nevada, some individuals migrate downslope to winter habitats (Verner
et al.
1992b). Laymon (1988) observed the subspecies migrating from summer home ranges in mixed conifer forests to winter home ranges in lower elevation pine-oak woodlands. He believed that similar migrations may also occur in Southern California. Tibstra (1999) observed that 10 of 22 dispersing juvenile owls having natal sites in coniferous forest habitats above 1,120 meters (m) (3,675 feet (ft)) moved downslope to lower elevation (305 m (1,000 ft) to 732 m (2,402 ft)) pine-oak woodland habitats. Of those ten, data were available through the following spring for only two, both of which overwintered and then moved back to high-elevation sites. The elevational movements of those two owls were significantly correlated with environmental temperature. Tibstra speculated that the pattern of migration to winter range observed in some adults may be established in the first year by dispersing juveniles.

Owls that migrate downslope do so between early October and mid-December and return in late February to late March. Such migrations range from 15 to 65 kilometers (km) (9 to 40 miles (mi)) with altitudinal changes of 500 to 1,500 m (1,640 to 4,921 ft). Some individuals have also been observed to move between high- and low-elevation ranges one or more times within a single winter (Gutiérrez
et al.
1995).

Individuals of migratory pairs of California spotted owls migrate to separate winter ranges rather than wintering together. After they return to their summer ranges, they follow the same breeding cycle as nonmigratory pairs, as described below. However, they probably do not spend as much time together at the beginning of the breeding season, because they may not return from their winter range by the time nonmigratory pairs have begun roosting together (Verner
et al.
1992b, Gutiérrez
et al.
1995). Individuals of nonmigratory pairs of California spotted owls remain together on the same home range year round, but they do not usually roost together during the winter. However, late in the winter, they increasingly roost together, preen each other, and occasionally copulate. For approximately two weeks before the first egg is laid, pairs roost together and copulate once or twice each evening. For about one week before the first egg is laid, the female spends most of her time near the nest, and the male brings her prey items (Verner
et al.
1992b, Gutiérrez
et al.
1995).

California spotted owl eggs are elliptical, white to pearl grey, and smooth to slightly granular in texture. California spotted owl egg laying peaks in mid-April. When egg laying begins, the female spends almost all her time in the nest, and the male supplies almost all of her food. The number of eggs in clutches ranges from one to four, with most nests containing two. Successive eggs are laid approximately three days apart. Pairs continue to copulate throughout, and for up to four days, after the egg laying period (Verner
et al.
1992b, Gutiérrez
et al.
1995).

Only the female incubates the eggs. During the first two days of incubation, she may leave the nest for up to two hours, but thereafter she will only leave the nest for 10 to 20 minutes at a time to regurgitate pellets, defecate, preen, or accept food from her mate (Verner
et al.
1992b).

Eggs hatch after approximately 30 days. Hatchlings are covered with white natal down, with juvenile plumage starting to replace natal down at about 10 to 20 days (Gutiérrez
et al.
1995). The female broods the hatchlings almost continuously for eight to ten days. During this period, the male supplies food for the female and young. Two to three weeks after the eggs hatch, the female begins foraging for one to four hours per night. Males have not been observed to feed the chicks directly, but continue to bring food to the nest, which the female passes to the chicks (Verner
et al.
1992b).

Most chicks fledge 34 to 36 days after hatching. New fledglings are weak fliers and may spend hours or days on the ground. Approximately three days after fledging, most young are able to fly or climb to elevated perches. Within a week, most are able to fly between trees. Both parents continue to feed the fledglings until mid to late September (Verner
et al.
1992, Gutiérrez
et al.
1995).

Dispersal.
Spotted owls primarily disperse as juveniles (natal dispersal), but may also disperse as adults (breeding dispersal) if habitat within their home range has been degraded or if they have separated from a mate (Verner
et al.
1992b). Natal dispersal occurs in September and October.

Natal dispersal distances of California spotted owls have been estimated using radio telemetry (Verner
et al.
1992, Tibstra 1999) and recapturing territorial owls that were banded as juveniles (LaHaye
et al.
2001, Jennifer Blakesley, Colorado State University,
in litt.
2002a). Dispersal distances of successfully dispersing owls ranged from 3 km (2 mi) to 76 km (47 mi). Mean natal dispersal distance of 26 owls in the Sierra National Forest and Sequoia National Park estimated using radio telemetry was 15.9 km (9.9 mi) (Tibstra 1999) and median distance of 42 owls on the Lassen National Forest estimated using recapture data was 25 km (16 mi) for females and 23 km (14 mi) for males (Blakesley
in litt.
2002a). Mean natal dispersal distances of 129 owls in southern California estimated using recapture data were 10.1 km (6.3 mi) for males and 11.7 km (7.3 mi) for females. No significant difference existed in dispersal distance or time to become territorial between sexes (LaHaye
et al.
2001). In this study, some dispersing owls did not occupy territories until they were four years old, but over 60 percent occupied territories within one year of fledging. Apparent survival of fledglings (calculated as the percentage of banded fledglings that were later relocated) was 31.8 percent.

LaHaye
et al.
(2001) concluded that the presence of conspecifics (members of the same species) may play a vital role in the recruitment of dispersing California spotted owls into a territory, because owls that “settled” (established territories) were significantly more likely to do so in territories that were occupied the previous year than would

be expected by chance and all previously vacant territories that were settled were adjacent to occupied territories. The percentage of territories occupied varied from 59 to 95 percent from year to year. During the study, young fledged from 28 percent of the 39 territories that were “frequently vacant,” indicating that habitat at those sites was suitable to support California spotted owl reproduction.

Four color banded adults on the Sierra National Forest later shifted territories, moving 3.4 km (2.1 mi), 3.5 km (2.2 mi), 3.9 km (2.4 mi), and 7.1 km (4.4 mi) (Verner
et al.
1992b). In a study of breeding dispersal of California spotted owls in the San Bernardino Mountains (LaHaye and Gutiérrez
in litt.
2002), 46 females and 38 males dispersed, which were 22 percent and 17 percent of the total banded females and males, respectively. Among dispersing females, 70 percent were adults and 30 percent subadults; among males, 71 percent were adults and 29 percent were subadults. A significantly higher percentage of subadults dispersed (30 percent) compared to the territorial population as a whole (14 percent). Mean dispersal distances were 4.3 km (2.7 mi) for females and 3.0 km (1.9 mi) for males, which are significantly shorter than natal dispersal distances observed in the same population.

Interactions with Other Species and Natural Mortality.
Spotted owls are mobbed by many species of diurnal birds (Gutie

rrez
et al.
1995). Red-tailed hawks (
Buteo jamaicensis
) and common ravens (
Corvus corax
) may take away prey items that are captured by spotted owls. The spotted owl's closest competitors are great horned owls (
Bubo virginianus
) and barred owls (
Strix varia
). Barred owls have recently colonized portions of the range of California spotted owls and are known to displace spotted owls from their territories (Verner
et al.
1992b, Gutiérrez
et al.
1995). Circumstantial evidence suggests that barred owls may kill spotted owls (Leskiw and Gutiérrez 1998). Northern goshawks (
Accipiter gentilis
), great horned owls, red-tailed hawks and potentially other birds of prey eat spotted owls (Verner
et al.
1992, Gutiérrez
et al.
1995). Fishers (
Martes pennanti
) have been observed in spotted owl nest trees and may take eggs or chicks (Gutiérrez
et al.
1995).

Starvation (Verner
et al.
1992b, Gutiérrez
et al.
1995, Tibstra 1999) has been documented as a cause of death in California spotted owls. Starvation is more common in juveniles than adults and may result from low prey availability or lack of hunting experience (Verner
et al.
1992b). Dispersing juveniles sometimes roost in open habitats during inclement weather, which may result in exposure causing or contributing to their deaths (Gutiérrez
et al.
1995). Accidents leading to death have been documented for spotted owls, including flying into obstacles and drowning (Verner
et al.
1992b).

Feeding and Metabolism.
Spotted owls are “perch and pounce” predators, hunting primarily by selecting an elevated perch, detecting prey by sight or sound, and swooping from the perch to capture the prey with their talons. Spotted owls are not fast fliers, but they are very agile and maneuverable. The flight pattern is a series of quick wing beats interspersed with gliding flight. Spotted owls use gliding flight when approaching prey. When gaining altitude in the forest canopy, they make a series of short climbing flights rather than one continuous flight. Flight is labored when attempting to fly to a higher perch or a nest sight. Flight above the forest canopy is probably rare, except during dispersal (Gutiérrez
et al.
1995). If a potential prey item is inaccessible or at a considerable distance from an owl's perch, the owl may move closer before pouncing (Verner
et al.
1992b). Spotted owls will forage at several sites within a single night (Gutiérrez
et al.
1995). They also hunt by capturing in mid-air flying prey such as insects, bats, and birds (Verner
et al.
1992b, Gutiérrez
et al.
1995). California spotted owls forage primarily at night, but have been observed hunting during the day, especially while raising young (Laymon 1991, Verner
et al.
1992). They may cache prey items on limbs, stumps, or the ground for later consumption (Gutiérrez
et al.
1995). Prey items include mammals, birds, and insects.

Spotted owls have a high water need relative to their metabolic rate (Weathers
et al.
2001), and have been observed drinking surface water from seeps and creeks (Gutiérrez
et al.
1995). California spotted owls have a narrow thermal neutral zone (the ambient temperature range through which a bird or mammal can maintain its normal body temperature without expending energy to do so) relative to birds in general and are therefore especially subject to heat stress (Gutiérrez
et al.
1995, Weathers
et al.
2001). They roost higher in the forest canopy during winter and lower during the summer. They will also move during a day in response to changes in ambient temperature and sun exposure. The variety of microclimates available in mature and old growth forests has been postulated as an explanation for the spotted owl's use of such habitats (Gutiérrez
et al.
1995).

Distribution, Range, and Land Ownership

Grinnell and Miller (1944) described the range of the California spotted owl as “[I]n general, coastal slope of southern California from southern San Diego County northwest to Santa Barbara, Ventura, and western Kern Counties, and west flank of Sierra Nevada north from Tulare County to Tehama County” and noted that the southern California range was apparently separated from the Sierra portion of the range.

The mapped range of the California subspecies in Grinnell and Miller (1944) indicated a gap in the distribution of spotted owls in Shasta County, separating the California and northern spotted owl subspecies. However, based on newer records and the occurrence of apparently-suitable habitat in the area, more recent authors have concluded that this purported gap between the California and northern subspecies may not have actually existed (Detrich
et al.
1993). For regulatory purposes, we established the “Pit River area” as the boundary between the northern spotted owl and the California spotted owl (55 FR 26114). No historic data are available regarding pre-European settlement population numbers of the California spotted owl.

The northern spotted owl ranges from southwestern British Columbia, Canada through western Washington, western Oregon, and northern California south along the coast to San Francisco Bay (Service 1990). The range of the Mexican spotted owl is disjunct from the other subspecies, from southern Utah and Colorado south through Arizona and New Mexico, and is discontinuous through the Sierra Madre Occidental and Oriental to the mountains at the southern end of the Mexican Plateau (Service 1993).

Today the California spotted owl still occurs throughout its historic range, including the west side of the Sierra Nevada from Shasta County south to the Tehachapi Pass, and all major mountains of southern California, including the San Bernardino, San Gabriel, Tehachapi, north and south Santa Lucia, Santa Ana, Liebre/Sawmill, San Diego, San Jacinto, and Los Padres ranges (Beck and Gould 1992). In addition, a few sites have been found on the eastern side of the Sierra Nevada and in the central Coast Ranges at least as far north as Monterey County.

Regarding the current distribution of the California spotted owl, Verner
et al.
(1992a) stated “in spite of the fact that

logging has occurred over nearly all of the conifer forests of the Sierra Nevada in the past 100 years, and especially the past 50 years, spotted owls continue to be widely distributed throughout most of the conifer zone. Indeed, spotted owls may be more abundant in some areas of the Sierra Nevada today than they were 100 years ago, “ (due to presumed effects of 19th century sheep grazing on spotted owl prey species.) They also stated that “Spotted owl distribution in the Sierra Nevada is characterized by its continuity and relatively uniform density.”

The elevation of known nest sites of California spotted owls ranges from about 305 to 2,348 m (1,000 to 7,700 ft), with approximately 86 percent of sites occurring between 915 and 2,135 m (3,000 and 7,000 ft) (USFS 2001a). In conifer forests mean elevation of nest sites was 1,160 m (5,300 ft) in the northern Sierra Nevada and 1,830 m (6,000 ft) in southern California (Gutiérrez
et al.
1992).

The California Department of Fish and Game (CDFG) has maintained a database of the number and location of California spotted owl territories located from the early 1970s to the present. We have combined that database with similar data collected by Sierra Pacific Industries, the major private timberland owner in the Sierra Nevada. The following discussion of locations and land ownership is based on that combined database and includes all records available to us. It is important to note that not all territories are occupied during any given year. The data presented are useful to illustrate the range of the species and jurisdictions under which it occurs, but should not be viewed as a population estimate because the current status of many territories is unknown due to lack of recent surveys; not all territories are occupied in a given year; and, in addition to territorial owls that comprise most of the sites in the database, nonterritorial, “floater” owls may be present but uncounted.

California spotted owl territories have been located on Forest Service (USFS), National Park Service (NPS), Bureau of Land Management (BLM), California Department of Parks and Recreation (State parks), California Department of Forestry and Fire Protection (CDF), California State Lands Commission (CSLC), Native American and private lands, and in Mexico.

Sierra Nevada.
In the Sierra Nevada the California spotted owl is mostly continuously and uniformly distributed, with several breaks in distribution where habitat appears limited due to natural or human-caused factors (Beck and Gould 1992). These Areas of Concern are further discussed in a later section.

In Sierra Nevada national forests, 99 percent of owl sites occur on the Lassen, Plumas, Tahoe, Eldorado, Stanislaus, Sierra, and Sequoia National Forests. The number of territories per national forest are as follows: Modoc 3, Lassen 138, Toiyabe 2, Inyo 5, Tahoe 173, Lake Tahoe Basin Management Unit 14, Plumas 254, El Dorado 202, Stanislaus 234, Sierra 226, and Sequoia 148. This results in a sub-total for Forest Service Sierran lands of 1,399 sites. The number of territories per national park are as follows: Lassen 6, Kings Canyon 19, Sequoia 50, and Yosemite 54. Fourteen territories are on BLM land in the Sierra Nevada. Three territories are on State parks, 1 is on CDF land, and 4 are on CSLC land. One territory is on Native American land and 314 are on private lands. Thus, the total number of California spotted owl sites known in the Sierra Nevada is 1,865 (Service 2002)

Because the subspecies has large home ranges, a given home range may occur across different ownerships. For instance, the Forest Service reported that over 15 percent of 135 Forest Service sites analyzed had greater than 15 percent of their theoretical home range on private lands (USFS 2001).

Coast Ranges and Southern California.
In southern California, the owl occupies “islands” of high elevation forests isolated by lowlands covered by chaparral, desert scrub, and increasingly (Noon and McKelvey 1992), human development (LaHaye
et al.
1994). California spotted owls have been found on 440 territories or sites in southern California in 15 populations comprised of 3 to 270 individuals, and separated from each other by 10 to 72 km (6 to 45 mi) (Verner
et al.
1992a, Gutiérrez 1994). Seventy-five percent of known territories are on Federal lands and twenty-five percent are on nonFederal lands. The Angeles National Forest has 64 territories, Cleveland National Forest has 18 territories, Los Padres National Forest has 109 territories, and San Bernardino National Forest has 138 territories; two territories are on BLM land; eight territories are on State parks; six are on Native American lands, 95 are on private lands, and one is in Mexico.

Within the California coastal and inland mountain ranges where California spotted owls occur (San Bernardo, San Gabriel, San Jacinto, Castaic, Santa Ana, and Santa Lucia mountains and San Diego/Peninsular, and Los Padres Ranges ) an area of just over 2,428,068 hectares (ha) (6,000,000 acres (ac)) was assessed for all habitats by the Forest Service (Stephenson and Calcarone 1999). Landownership in the assessment area is National Forest (57 percent), private (33 percent), BLM (4 percent), Indian (3 percent), State (2 percent), military (1 percent) and local (1 percent). Not all of the analysis area is suitable spotted owl habitat (mixed conifer hardwood), thus the portion of the total owl population or sites known on Federal lands as determined in Verner
et al.
1992a and Gutiérrez 1994, is higher (75 percent) than their relative ownership in the assessment area (62 percent).

The range of California spotted owls in southern California is disjunct from that in the Sierra Nevada range as a result of natural topographic and manmade factors (Stephenson and Calcarone 1999). Within this southern range, habitat and spotted owls are distributed discontinuously across the landscape reflecting natural vegetation breaks, topographic conditions, and human induced habitat disturbance and fragmentation (Noon and McKelvey 1992). The spotted owls in the southern portion of the range may function as a meta population, with separate subpopulations connected by infrequent but persistent interchange of individual owls (Noon and McKelvey 1992; LaHaye
et al.
1994).

Habitat

The habitat used by California spotted owls today is comprised of forests that have been shaped by numerous inter-acting human impacts, including timber harvest, livestock grazing, urbanization, and, because of fire suppression, changes in the character of wildfires. Prior to the occupation of California by Anglo-Americans in the mid-1900s, habitat was probably fairly stable on a large geographic scale, although there were almost certainly localized variations caused by fire and other causes of forest mortality. In recent decades, timber harvest and ingrowth related to fire suppression have created widespread forest conditions believed to be considerably different than that of pre-historic times (McKelvey and Weatherspoon 1992, McKelvey and Johnston 1992). In the following section, the current understanding of use of today's forests by California spotted owls will be portrayed, along with some discussion of the factors that created these conditions. The anticipated trends in habitat will be discussed in the Threats section below.

The suppression of wildfire during the 20th century has been one of the most important factors in creating the forest conditions that provide habitat for

the California spotted owl today. For thousands of years preceding European settlement of California, low to moderate intensity fires burned frequently in most Sierra Nevada vegetation types (University of California 1996). Median fire return intervals were typically less than 20 years, and as low as four years, in ponderosa pine (
Pinus ponderosa
) and mixed conifer zones. In the mixed conifer zone, where approximately 80 percent of Sierra Nevada California spotted owl sites occur (see Habitat Requirements, above), many plant species take advantage of, or depend on, fire for their reproduction or as a means of competing with other species. The effects of frequent surface fires largely explain the reports and photographs by early observers who described Sierra Nevada forests as typically open and park-like. However, other early observers reported dense conditions and dark or impenetrable forest. These records suggest that although open conditions were more prominent than they are today, Sierra Nevada forests were a mix of different degrees of openness, with an unknown proportion in dark, dense, nearly impenetrable vegetative cover and with variations in density with latitude, aspect, and elevation (University of California 1996, Gruell 2001).

Suppression of wildland fires had been established in California as State and Federal policy by the early 20th century. The area burned annually in recent times has been estimated to be only about three percent of that burned pre-European settlement in mixed conifer forest types (University of California 1996). As will be discussed further below, fire suppression has resulted in substantial growth of small understory trees in much of the range of the California spotted owl.

Timber harvest has been another obvious impact to California spotted owl habitat (Gutiérrez 1994, Verner
et al.
1992a). McKelvey and Johnston (1992) used historical documents to describe the status of Sierran forests at the beginning of the 20th century, and detailed the harvest history from the late 19th century to 1990. Harvest steadily intensified from the railroad building and mining eras of the 1800s until the 1950s, then remained at relatively high levels through the 1980s. (Intermittent declines occurred during poor economic conditions of the 1930s and early 1980s.) Low elevations and accessible areas (McKelvey and Johnston 1992, Beardsley
et al.
1999) and commercially important forest types such as west-side mixed conifer and east-side pine (Franklin and Fites-Kaufmann 1996) have been the most heavily impacted. As a result, McKelvey and Johnston stated that “The mixed conifer zone of the Sierra Nevada * * * has few or no stands remaining that can be described as natural or pristine.”

Verner
et al.
(1992a) discussed five major factors of concern for California spotted owl habitat that have resulted from historical timber harvest strategies: (1) Decline in the abundance of very large, old trees; (2) decline in snag density; (3) decline in large-diameter logs; (4) disturbance or removal of duff and topsoil layers; and (5) change in the composition of tree species. Of these concerns, they believed significant changes in diameter distributions of trees in the Sierra Nevada and rapid reductions in the distribution and abundance of large, old, and decadent trees posed the greatest threat to the California spotted owl. Thus, extensive commercial harvest of large old trees in late successional forest directly affected the key structural components of California spotted owl habitat.

Timber harvest in the Sierra Nevada peaked in the 1950s and remained at high levels into the late 1980s (McKelvey and Johnston 1992). Since the late 1980s, the volume of timber harvested in the Sierra Nevada has declined substantially. In particular, levels of timber harvest on national forest lands declined after implementation of the California Spotted Owl Sierran Province Interim Guidelines in 1993 (USFS 2001a). From Fiscal Year (FY)1998 through FY 2002, the mean annual total harvest volume (279.4 million board ft (mmbf)) on the seven National Forests that support most of the Sierra owl population was about 28 percent of the mean annual total volume harvested on those forests during the period FY 1986 through FY 1990 (1,007 mmbf) (USFS 2002a). Whereas old-growth accounted for most of the volume in the past, more recent harvest practices have focused on thinning of young, smaller trees (McKelvey and Johnston 1992).

The decline in Federal harvest led an overall decline in the total Federal and private harvest in the Sierra Nevada. According to California timber tax data (California Board of Equalization 2002), total harvest from public lands in 18 Sierra Nevada counties during the late 1980s and early 1990s constituted about half of the total annual volume harvested in those counties, but following the 1993 implementation of protections for the California spotted owl on Forest Service lands, the public lands harvest did not exceed 25 percent of the total annual volume harvested from those counties in any year from 1994 through 2000. In the meantime, private harvest during the 15-year period from 1986 to 2000 remained between 650 and 775 mmbf per year, except for a 2-year spike of over 900 mmbf per year in 1990 and 1991. Mean annual volume from private lands in the 18 Sierra Nevada Counties in the period for the period 1986 to 1990 was about 811 mmbf, and mean annual volume from 1996 to 2000 was about 714 mmbf, a difference of 12 percent. Thus, in the Sierra Nevada, private lands harvest has declined somewhat while Federal harvest has declined sharply since the late 1980s.

Similar trends in timber harvest have occurred in the four southern California national forests, although timber harvest in this area was never as extensive as in the Sierra Nevada. According to McKelvey and Johnston (1992), harvest volume in Los Angeles and San Bernardino counties was about ten to twenty times higher in the 1960s than in the early 1980s, and the decline has continued since the 1980s. Southern California national forests have not had a commercial green timber sale program for over a decade. Harvest in recent years has primarily been salvage and hazard trees along roads and near administrative sites (M. Gertsch, USFS, pers. comm. 2002). Mean annual total harvest volume for the four forests in FY 1998 to 2002 (1.66 mmbf) was about 30 percent of the mean annual total harvested on the four forests during FY 1988 to1992 (5.48 mmbf) (USFS 2002a).

Thus, timber harvest, the primary cause of habitat loss for the California spotted owls for decades, has been much reduced in recent years. Spotted owls today are occupying habitat that is a combination of the remnants of older stands and stands regenerating from timber harvest in past decades. The present habitat used by California spotted owls is further described below.

California spotted owls use a broader range of habitat types than the northern spotted owl (Call
et al.
1992, Gutiérrez
et al.
1992, Anderson and Mahato 1995, Moen and Gutiérrez 1997, North
et al.
2000), in part due to the relatively more complex landscapes available to the California subspecies (Zabel
et al.
1992b, Franklin and Fites-Kaufmann 1996, Helms and Tappeiner 1996, Beardsley
et al.
1999). In the Sierra Nevada, this complexity reflects: (1) The variety of environmental conditions due to elevation, latitude, geology, precipitation, and temperature; (2) rich flora; and (3) influence of natural disturbance, especially fire (Andersen and Mahato 1995) and human disturbance (Franklin and Fites-Kaufmann 1996). The forests of the

Sierra Nevada have a complex logging history dominated by selection methods (McKelvey and Johnston 1992, Beardsley
et al.
1999) varying by number of entries, types of species harvested, size distribution of harvested trees, and total volume logged (Zabel
et al.
1992b). The heterogeneity of forests occupied by California spotted owls make quantifying its habitat difficult and sensitive to scale. Several studies have found that analysis of habitat at a coarse, small scale (
e.g.
, using timber type polygons developed for timber management) masks fine grained attributes used or selected by owls (Bias and Gutiérrez 1992, Zabel
et al.
1992a, Moen and Gutiérrez 1997).

Despite the complexity of California spotted owl habitat, several authors have concluded the subspecies is a habitat specialist (Andersen and Mahato 1995, Moen and Gutiérrez 1997, LaHaye
et al.
1997), selecting habitat at several scales. California spotted owls, like the other subspecies of spotted owls, use or select habitats for nesting, roosting, or foraging that have structural components of old forests, including large (typically greater than 61 cm (24 in) diameter at breast height (dbh; breast height has been standardized at 137 cm (4.5 ft) above the ground) (Call 1990, Gutiérrez
et al.
1992, Zabel
et al.
1992a, Moen and Gutiérrez 1997, USFS 2001a), decadent trees (trees with cavities, broken tops, etc.); high density of trees (Laymon 1988, Call 1990, Bias and Gutiérrez 1992, Gutiérrez
et al.
1992, LaHaye
et al.
1997, Moen and Gutirrez 1997); multi-layered canopy/complex structure (Call 1990, Gutiérrez
et al.
1992, LaHaye
et al.
1997, Moen and Gutiérrez 1997); high canopy cover (greater than 40 percent and mostly greater than 70 percent; Laymon 1988, Bias and Gutiérrez 1992, LaHaye
et al.
1992, Gutiérrez
et al.
1992, Zabel
et al.
1992a, Moen and Gutiérrez 1997, North
et al.
2000); snags (Laymon 1988, Call 1990, Bias and Gutiérrez 1992, Gutiérrez
et al.
1992, LaHaye
et al.
1997 ); and logs (Call 1990). Gutiérrez
et al.
(1992) noted that these characteristics applied to mixed conifer forests, because riparian/hardwood forests occupied by California spotted owls did not necessarily have these characteristics.

Late successional forests provide habitat attributes selected by California spotted owls, including large trees, high canopy closure, multi-layered canopies, snags, and logs (University of California 1996). The current extent of old forests in the Sierra Nevada is believed to be substantially less than in pre-historic times. Estimates of the current extent have been made by several authors. The University of California (1996) reported that in national parks in the Sierra Nevada, which contain the best representation of pre-European settlement conditions because only minor areas have been subject to timber harvest, 55 percent of forests are in late successional conditions, but on all Federal lands in the Sierra Nevada, late successional conditions are now found on only 19 percent of forest lands. The Forest Service (USFS 2001a) reported that old forest conditions have declined from 50 to 90 percent in various vegetation types compared to the range of historical conditions. Beardsley
et al.
(1999) estimated that approximately 15 percent of coniferous forests in the Sierra Nevada remain in high quality old growth/late successional stages; most of these stands are in high elevations and national parks (Franklin and Fites-Kaufmann 1996). Most of the remaining high quality late successional/old growth habitat in the Sierra Nevada is in public ownership; less than two percent of 1,214,000 ha (3 million ac) of private land was classified as high quality late successional/old growth habitat (Franklin and Fites-Kaufmann 1996).

California spotted owls in the Sierra Nevada may have undergone at least three periods of decline: (1) Elimination of prey species by intensive livestock grazing and burning in the 1800s; (2) logging beginning in the late 1800s, which removed basic structural elements of owl habitat; and (3) recent

California spotted owls in the Sierra Nevada may have undergone at least three periods of decline: (1) Elimination of prey species by intensive livestock grazing and burning in the 1800s; (2) logging beginning in the late 1800s, which removed basic structural elements of owl habitat; and (3) recent logging of stands that regenerated following initial entry (Gutiérrez 1994).

In the early 1990s, researchers expressed concern regarding potential lag effects in population decline, in which the negative effects of habitat modification might not be observed until subsequent years (Noon
et al.
1992, Gutiérrez 1994, LaHaye
et al.
1994). However, it seems reasonable to presume that the causal mechanisms of negative effects ascribed to the high levels of timber harvest circa 1990 were substantially reduced as timber harvest levels dropped and increased protection measures were instituted in the mid- and late-1990s. In the opinion of the Forest Service (USFS 2001a), it is unlikely that recent timber harvest on national forests has caused declines in spotted owl populations, although the possibility exists that declines are due in part to latent effects of past timber harvest. However, because the regeneration of habitat may take several decades, modification of habitat components that resulted from past timber harvest is probably still affecting the subspecies to various degrees.

Although late-successional forests with large old trees are believed to provide the best habitat for California spotted owls, descriptions of suitable habitat derived from habitat use studies often also include smaller size classes, and thus, include a greater proportion of the landscape than that included in the above estimates of older forest extent. In the early 1990s, Verner
et al.
(1992a) estimated the amount of suitable California spotted owl habitat on public land. They defined suitable habitat as having canopy cover exceeding 40 percent and dominant trees 30 to 36 cm (12 to 14 in) dbh or larger, or areas with the potential to reach those values relatively rapidly. Their estimates of California spotted owl habitat in the Sierra Nevada by jurisdiction were as follows: 1,416,400 ha (3,500,000 ac) on national forests;186,560 ha (461,000 ac) on national parks; 27,721 ha (68,500 ac) on BLM lands; and 10,522 ha (26,000 ac) on State lands. Their estimates for the Coast ranges and southern California were: 218,530 ha (540,000 ac) on national forests; 3,076 ha (7,600 ac) on BLM; and 10,117 ha (25,000 ac) held by State and local governments. Thus, the total estimated habitat was 1,872,926 ha (4,628,100 ac). An undetermined amount of suitable habitat also existed on private and Native American lands (Verner
et al.
1992a).

Habitat in the Sierra Nevada.
The Final Environmental Impact Statement (FEIS) for the Sierra Nevada Framework Amendment (SNFPA) (USFS 2001a) estimated the amount of suitable habitat for California spotted owls on national forest lands in the Sierra Nevada to be 1.7 million ha (4.3 million ac). This estimate was about 14 percent higher than that of Verner
et al.
(1992a). The new estimate was based on more refined analysis, rather than an actual increase in habitat. This constitutes about 59 percent of the forested lands on the Sierra Nevada national forests.

Amounts of habitat on private lands have not been quantified. Generally, industrial landowners regard information relevant to timber inventories as proprietary. Based on Forest Service data, there were about 485,600 ha (1.2 million ac) of industrial timberland in the Sierra Nevada as of 1994 (derived from Waddell and Bassett 1997a, Waddell and Bassett 1997b, Waddell and Bassett 1997c). National forests in the Sierra Nevada include

approximately 560,000 ha (1.4 million ac) of private land within their administrative boundaries. Private land inholdings are much greater in extent in the northern national forests (especially the Lassen, Plumas, and Tahoe) than in the southern Sierra Nevada forests. Much of the private land within the boundary of the Lassen and Plumas National Forests is in contiguous blocks, leaving national forest lands also fairly contiguous. Most private land on the Tahoe National Forest is in checkerboard ownership, and the Eldorado National Forest has a combination of checkerboard ownership and large contiguous blocks of inholdings. We acknowledge that considerable amounts of suitable habitat exist on private lands, especially in the smaller size classes. This is reflected in the occurrence of over 300 spotted owl activity centers (about 17 percent of the Sierra Nevada total) on private lands.

The mixed-conifer forest type (sugar pine (
Pinus lambertiana
), ponderosa pine, white fir (
Abies concolor
), Douglas-fir (
Pinus lambertiana
), giant sequoia (
Sequoiadendron giganteum
), incense-cedar (
Calocedrus decurrens
), black oak (
Q. kelloggii
), and red fir (
Abies magnifica
) is the predominant type used by spotted owls in the Sierra Nevada: about 80 percent of known sites are found in mixed-conifer forest, 10 percent in red fir forest (red and white fir, lodgepole pine (
Pinus contorta
), and quaking aspen (
Populus tremuloides
), seven percent in ponderosa pine/hardwood forest type (ponderosa pine, interior live oak (
Quercus wislizenii
), canyon live oak (
Quercus chrysolepis
), black oak, incense-cedar, white fir, tanoak (
Lithocaarpus densiflorus
), and Pacific madrone (
Arbutus menziesii
)), and the remaining three percent in foothill riparian/hardwood forest (cottonwood (
Populus
ssp.), California sycamore (
Platanus racemosa
), interior live oak, Oregon ash (
Fraxinus latifolia
), and California buckeye (
Aesculus californica
) and east-side pine (ponderosa and Jeffrey pine (
P. jeffreyi
)) (Verner
et al.
1992a, USFS) 2001).

Six major studies (Guti

errez
et al.
1992) have described habitat relations of the owl in four areas spanning the length of the Sierra Nevada. These studies examined spotted owl habitat use at three scales: landscape; home range; and nest, roost, or foraging stand. Based on comparisons of time spent by owls in various habitat types to amounts of habitat available, owls preferentially use areas with at least 70 percent canopy cover, use habitats with 40 to 69 percent canopy cover in proportion to their availability, and spend less time in areas with less than 40 percent canopy cover than might be expected if habitat were selected randomly.

California spotted owls in the Sierra Nevada prefer stands with significantly greater canopy cover, total live tree basal area, basal area of hardwoods and conifers, and snag basal area for nesting and roosting. Owls use stands dominated by trees with dbhs between 30 and 61 cm (12 and 24 in) and canopy covers between 40 and 100 percent for nesting significantly more than expected, based on the proportion of those forest types (Guti

errez
et al.
1992). Stands suitable for nesting and roosting have: (1) Two or more canopy layers; (2) dominant and codominant trees in the canopy averaging at least 61 cm (24 in) in dbh; (3) at least 70 percent total canopy cover (including the hardwood component); (4) higher than average levels of very large, old trees; and (5) higher than average levels of snags and downed woody material (Guti

errez
et al.
1992, USFS 2001a).

Analysis of vegetation characteristics of plots surrounding 292 California spotted owl nest and roost sites on the Lassen, Eldorado, and Sierra National Forests, and in Sequoia and Kings Canyon National Parks provides further information on habitat types favored by the species (USFS 2001a). Thirty-two percent of the plots were in stands with multilayered canopies exceeding 60 percent closure and with average trees exceeding 61 cm (24 in) in dbh. Eighteen percent were in stands with 40 to 59 percent canopy closure and with average trees exceeding 61 cm (24 in) in dbh. Fourteen percent were in stands with over 60 percent canopy cover and with average trees between 28 and 61 cm (11 to 24 in) in dbh. Eleven percent were in stands with 40 to 59 percent canopy closure and average trees between 28 and 61 cm (11 to 24 in) in dbh. Nine percent were in stands with over 60 percent canopy closure and average trees exceeding 61 cm (24 in) in dbh. Seven percent were in stands with 25 to 39 percent canopy cover and average trees exceeding 61 cm (24 in) in dbh. Five percent were in stands with 25 to 39 percent canopy cover and average trees between 28 and 61 cm (11 to 24 in) in dbh. North
et al.
(2000) suggested that canopy cover, tree density, and foliage volume represent conditions consistent across different forest types and therefore could indicate the basic nest site conditions selected by California spotted owls. California spotted owl nests were consistently located in sites with 75 percent canopy cover, 300 trees/ha (122 trees/ac), and 40,000 cubic m/ha (571,860 cubic ft/ac) of foliage volume.

Moen and Gutie

rrez (1997) analyzed California spotted owl habitat at the landscape, habitat patch, and microsite levels on a 355 square kilometer (137 square mile) study area on the El Dorado National Forest. They used remote sensing to analyze vegetation in 457 ha (1,129 ac) circular plots surrounding spotted owl activity centers, and compared those plots with randomly selected plots of equal size. Owl plots were significantly more homogeneous than random sites, indicating that owls select against patchy or fragmented habitats; owl sites contained significantly more area with canopy closure exceeding 70 percent than random plots; and California spotted owl roosts were significantly more likely to be located in mixed conifer habitat containing trees greater than 30 cm (12 in) dbh than would be expected by chance. In addition, of 82 roost sites examined, 56 (68 percent) were in habitat with greater than 40 percent canopy closure and trees greater than 30 cm (12 in) dbh, and 97 percent of roost sites had trees over 100 cm (39 in) dbh. Microsite comparison between sixteen 0.04 ha (0.10 ac) vegetation plots surrounding nest sites and random plots of equal area showed that nest plots had significantly higher structural diversity, more total trees, larger trees, and more trees over 100 cm (39 in) dbh.

Bias and Gutiérrez (1992) attributed low use of private timberlands by roosting and nesting California spotted owls to sanitation (removal of damaged or diseased trees or species of low commercial value) and high-grade logging (harvest of large trees of high commercial value) that removed potential nest trees. However, as stated above, California spotted owls do occur on private timberlands. Habitat use by California spotted owls has been studied on a private timber production area in the Sierra Nevada, 48 kilometers (km) (30 mi) east of Chico, California (Larry L. Irwin
et al.
, National Council for Air and Stream Improvement, Incorporated,
in litt.
2002). Seven pairs of California spotted owls were repeatedly located using radiotelemetry. Habitat use was similar to that observed in other studies on Federal lands. Owls were located in areas with canopy closure averaging 70 percent, dominated by trees 30 to 36 cm (12 to 14 in) in dbh but with a few larger (over 66 cm (26 in) dbh) trees, and with tree densities ranging from 930 to 1,360 trees/ha (372 to 544 trees/ac). To our knowledge, there are no studies providing information on demographic performance of owl populations on private lands in the range of the California spotted owl.

Habitat in the Coast Range and Southern California.
In the coast range, California spotted owls occupy redwood/California-laurel forests which consists of a mix of coast redwood (
Sequoia sempervirens
), California-laurel (
Umbellularia californica
), tanoak, Pacific madrone, red alder (
Alnus rubra
), and white alder (
Alnus rhombifolia
), coast live oak, Santa Lucia fir (
Abies bracteata
), and bigleaf maple (
Acer macrophyllum
) (Verner
et al.
1992a). Spotted owls can be found at elevations below 305 m (1,000 ft) along the Monterey coast to approximately 8,500 ft (2,591 m) in the inland mountains (Stephenson and Calcarone 1999). Lower elevation (less than 3,000 ft (914 m)) birds can be found in pure oak stands and higher elevation (greater than 6,500 ft (1,981 m)) birds can be found in pure conifer stands.

Verner
et al.
(1992a) noted that California spotted owls also use riparian hardwood forest types (coast and canyon live oak, cottonwood, California sycamore, white alder, and California laurel) in southern California. Owls on Mount San Jacinto used conifer and riparian hardwood forests significantly more than would be expected based on their availability and owls on Palomar Mountain primarily used conifer or mixed forests of conifers and hardwoods. California spotted owl nest sites in the San Bernardino Mountains were more likely to be located in areas with steeper slopes and in the lower third of canyons and owl nest and roost sites in this area were more likely to be located in areas with higher canopy closure and higher basal area (the area of all trees at breast height) than random sites.

Spatial positions and vegetation types were compared between plots surrounding 144 California spotted owl territory centers and 144 random plots in the San Bernardino Mountains of southern California (Smith
et al.
1999, Humboldt State University,
in litt.
2002). Owl sites were significantly closer to one another than random sites, showing a clumped distribution. Owl sites contained more area of closed canopy forest, larger mean patch sizes of closed canopy forest, and lower habitat diversity than random sites. California spotted owl territories in this study were found in three vegetation types; canyon live oak/ big cone Douglas-fir (39 percent of territories), mixed conifer/hardwood (which includes canyon live oak, big cone Douglas-fir, sugar pine, white fir, Coulter pine (
P. coulteri
), incense cedar, and black oak) (28 percent of territories), and mixed conifer (which contains white fir, Jeffrey pine, and incense cedar (33 percent of territories).

Stephenson and Calcarone (1999) estimated that there were approximately 473,473 ha (1,170,000 ac) of habitat types where spotted owls were known to reproduce (low-elevation oak/bigcone Douglas-fir, mid-elevation conifer/hardwood, and high elevation mixed conifer) within the range of the subspecies in southern California and the central Coast Ranges. The total amount of available suitable habitat in the analysis area is likely lower, because it is possible that not all habitat is currently in a condition suitable for reproduction, roosting or foraging.

Nest Tree Characteristics.
California spotted owls nest in a variety of tree/snag species in pre-existing structures such as cavities, broken top trees, and platforms such as mistletoe brooms, debris platforms and old raptor or squirrel nests (Gutie

rrez
et al.
1992, 1995). Nest trees are often large, over 89 cm (35 in) average dbh (Gutie

rrez
et al.
1992, Steger
et al.
1997, LaHaye
et al.
1997), and larger than other trees in the same stand (Gutie

rrez
et al.
1992). Nest trees are also often greater than 200 years old (Gutie

rrez
et al.
1992, North
et al.
2000). However, approximately 25 percent of nest trees out of a sample of over 250 were less than 76 cm (30 in) dbh (Gutie

rrez
et al.
1992). Although old, large trees are important to California spotted owls, intermediate-sized (28 to 61 cm (11 to 24 in)) trees were also selected by nesting (LaHaye
et al.
1997; and trees 51 to 76 cm (20 to 30 in) dbh), roosting (Moen and Gutie

rrez 1997), and foraging (Laymon 1988) owls.

Prey and Foraging Habitat.
California spotted owls are considered prey specialists (Verner
et al.
1992b) because they select a few key species (Verner
et al.
1992b) among the variety of taxa on which they prey. In the upper elevations of the Sierra Nevada (about 1,200 to 1,525 m (4,000 to 5,000 ft), the primary prey is the northern flying squirrel (
Glaucomys sabrinus
), which is most common in larger stands of mature forests (Verner
et al.
1992b). In lower elevations of the Sierra Nevada and in southern California, the primary prey is the dusky-footed woodrat (
Neotoma fuscipes
) (Thrailkill and Bias 1989), which is most abundant in shrubby habitats and uncommon in pure conifer forests or forests with little shrub understory (Williams
et al.
1992). Both flying squirrels and woodrats occur in the diets of owls in the central Sierra Nevada (Verner
et al.
1992b). Home ranges of owls in areas where the primary prey is northern flying squirrels are consistently larger than those where the primary prey is dusky-footed woodrats presumably because woodrats occur in greater densities and weigh more than flying squirrels (Zabel
et al.
1992a). Verner
et al.
(1992b) reported that approximately 25 percent of known owl sites in the Sierra Nevada occur where woodrats are the primary prey species and 75 percent of sites occur where flying squirrels are the primary prey species.

Other prey items include gophers (
Thomomys
spp.), mice (
Peromyscus
spp.), diurnal squirrels (
Tamiasciurus douglasii, Sciurus griseus, Spermophilus beecheyi, Eutamias
spp.) and a variety of other rodents; shrews (
Sorex
spp.); moles (
Scapanus
spp); bats (
Myotis
spp.); birds; frogs; lizards; and insects (Verner
et al.
1992b, Gutie

rrez
et al.
1995, Tibstra 1999). California spotted owls have low metabolic rates relative to other birds. Analysis of metabolic rates and the energy content of prey items indicates that an individual California spotted owl would need to eat one flying squirrel every 1.8 days or one woodrat every 3.7 days (Weathers
et al.
2001).

California spotted owls in the Sierra Nevada forage most commonly in intermediate- to late-successional forests with greater than 40 percent canopy cover and a mixture of tree sizes, some larger than 61 cm (24 in) in dbh. The birds consistently use stands with significantly greater canopy cover, total live tree basal area, basal area of hardwoods and conifers, snag basal area, and dead and downed wood than are found at random locations within the forest. Studies on the Tahoe and Eldorado National Forests found that owls forage in stands with large diameter trees (defined as trees greater than 61 cm (24 in) in dbh in one study and trees 51 to 89 cm (20 to 35 in) in dbh in the other) significantly more than expected based on availability. Owls also forage in stands with trees between 30 and 61 cm (12 and 24 in) dbh and greater than 70 percent canopy cover significantly more than expected, based on the proportion of that forest type (USFS 2001a).

Stands suitable for owl foraging have: (1) At least two canopy layers; (2) dominant and codominant trees in the canopy averaging at least 28 cm (11 in) in dbh; (3) at least 40 percent canopy cover in overstory trees (30 percent canopy cover in red fir dominated forests); and (4) higher than average numbers of snags and downed woody material. California spotted owls forage in forests with ample open flying space within and beneath the canopy (Gutie

rrez
et al.
1995); therefore, extremely dense stands may not be used

for foraging. Although canopy covers down to 40 percent are suitable for foraging, they appear to be so only marginally. Radio tracking data from the Sierra National Forest showed that owls tended to forage more in sites with greater than 50 percent canopy cover than predicted from their availability; while stands with 40 to 50 percent canopy cover were used about in proportion to their availability (USFS 2001a). The subspecies avoids open (0-30 percent canopy cover; Gutie

rrez
et al.
1992) or logged (Call 1990, Zabel
et al.
1992b, Gutie

rrez and Pritchard 1990) areas.

Winter Habitat.
Winter habitats of owls that undertake altitudinal migrations have similar canopy closures, but lower basal areas of both green trees and snags, and higher shrub densities than higher-elevation summer habitats (Gutie

rrez
et al.
1995).

General Description of Suitable Habitat.
Based on the above studies, nesting habitat for California spotted owls is generally described as stands with an average dominant and co-dominant tree diameter of greater than 24 in and canopy cover of greater than 70 percent. Foraging habitat is generally described as stands of trees of 30 cm (12 in) in diameter or greater, with canopy cover of 40 percent or greater. Exceptions to both descriptions are known to occur. Suitable habitat includes California Wildlife Habitat Relationship (WHR) habitat types 4M, 4D, 5M, 5D, and 6 (Mayer and Laudenslayer 1988).

Home Range

Spotted owl pairs have large home ranges that may overlap those of conspecifics (Verner
et al.
1992b). A portion of the home range is defended as a territory, especially against unknown intruders (Gutiérrez
et al.
1995). However, territorial disputes between neighbors are rare. Members of the same sex are more likely to display aggression toward each other than members of the opposite sex (Verner
et al.
1992b). Spotted owls may roost near conspecifics other than their mates (Gutiérrez
et al.
1995). Verner
et al.
1992b suggested that the spotted owl territorial system functions such that an individual or pair are dominant within a territory and prevent conspecifics from breeding there, but that feeding or roosting by those birds may be tolerated.

Carey
et al.
(1992) studied the relationship between the amount of habitat used by northern spotted owls and prey abundance within those habitats. They found that owls used more area in habitats where the estimated biomass of medium sized prey, primarily flying squirrels and woodrats, was lower. The largest home ranges of California spotted owls occur where flying squirrels comprise the majority of the owl's diet and the smallest occur where woodrats dominate (Verner
et al.
1992b, Zabel
et al.
1992a). Woodrat populations are denser than flying squirrel populations, often by at least 10 fold, and woodrats weigh nearly twice as much as flying squirrels. Variation in prey availability likely affects the percentage of California spotted owl pairs that nest and successfully fledge young. Weather may also affect these parameters, either by directly affecting the owls or by affecting their prey base (Verner
et al.
1992b).

Estimates of California spotted owl home range size are extremely variable. All available data indicate that they are smallest in habitats at relatively low elevations that are dominated by hardwoods, intermediate in size in conifer forests in the central Sierra Nevada, and largest in the true fir forests in the northern Sierra Nevada (Zabel
et al.
1992a, USFS 2001a). Based on an analysis of data from telemetry studies of California spotted owls, mean breeding season pair home range sizes have been estimated as 3,642 ha (9,000 ac) in true fir forests on the Lassen National Forest; 1,902 ha (4,700 ac) in mixed conifer forests on the Tahoe and Eldorado National Forests; and 1,012 ha (2,500 ac) in mixed conifer forests on the Sierra National Forest. Zimmerman
et al.
(2000) used radiotelemetry data to estimate the breeding season home range of two pairs of California spotted owls in the San Bernardino Mountains of southern California. The average home range (571 ha (1,410 ac)) was smaller than those reported for the Sierra Nevada and varied widely between the two pairs (325 to 816 ha (803 to 2,016 ac)).

Gutiérrez
et al.
(1992) analyzed the sizes of stands containing nest trees (
i.e.
, nest stands) and the cumulative sizes of each nest stand plus all adjoining stands that were in vegetation strata preferentially used by owls for nesting. The mean size of nest stands was about 40 ha (100 ac); the mean size of the nest stand plus adjacent suitable stands was about 120 ha (300 ac). In radio tracking studies, the central area including half of the foraging locations of owls was found to vary from an average of 128 ha (317 ac) on the Sierra National Forest to an average of 319 ha (788 ac) on the Lassen National Forest (Gutiérrez
et al.
1992). Bingham and Noon (1997) used radiotelemetry data to calculate core areas within the home ranges of four California spotted owls. Owls used the core areas more than would be expected if the entire home range were used at random. Core areas contained an average of 66 percent of points at which owls were located within an average of 21 percent of the home range.

Habitat in Home Range.
California spotted owls were found to select more consistently for habitat patches with high canopy cover than for large tree size-class (Zabel
et al.
1992a). Call (1990) estimated 42 percent of the home range to be medium timber 28 to 53 cm (11 to 21 in) dbh, and 55 percent large timber greater than 53 cm (21 in). The proportion of habitat in home ranges of owls in conifer forests of the Sierra Nevada with canopy cover greater than 40 percent was 68 percent and 81 percent for the two conifer sites studied. (Zabel
et al.
1992a).

California spotted owls have been known to use stands that were recently selectively harvested (Zabel
et al.
1992b). However, where forests in the Sierra National Forest were heavily thinned, owls consistently nested in patches with large, old, high crown-volume trees (North
et al.
2000), relying on the remaining components of the original forest.

Numerous studies have described habitat used by spotted owls and habitat that occurs around owl nest sites and activity centers, but the relationships between these forest habitat characteristics and the distribution and demographic performance of California spotted owls are not completely understood. Several studies that have related habitat characteristics with California spotted owl demographic performance and occupancy rates found that productivity was positively correlated with amounts of forest with high canopy cover. Blakesley (2002a) characterized habitat within 1,830 ha (4,532 ac) circles surrounding 67 California spotted owl nest sites in northeastern California and used those data to explain observed variation in site occupancy, apparent survival probability, reproductive output, and nest success. Site occupancy was positively associated with the amount of habitat dominated by large trees and high canopy cover. North
et al.
(2000) found higher reproduction in conifer forest associated with high foliage volumes and concluded: “The possible interaction of weather and nest-site structure on owl reproduction suggests forest managers should be cautious about reducing canopy volume in potential owl nesting areas. Retaining groups of large, old, high crown-volume trees may be needed to maintain the number of potential nesting sites in a forest.” Apparent survival and

reproductive output were positively correlated with the proportion of habitat surrounding each nest that was selected by the owls throughout the Sierra Nevada, as described by Gutiérrez
et al.
(1992). Nest success was positively associated with the presence of large trees within the nest stand.

Verner
et al.
(1992b) reported that about 75 percent of the California spotted owls in the Sierra Nevada occurred in areas where the northern flying squirrel was the primary prey species. Northern flying squirrels have been shown to be most common in larger stands of mature forests (Williams
et al.
1992). Flying squirrels typically use older mature forest because they provide suitable nest sites, including snags, and abundant sources of food including arboreal lichens and truffles, which are associated with an abundance of soil organic matter and decaying logs (Verner
et al.
1992b). In second-growth forests in Oregon, northern flying squirrels were found in younger forests if large snags and down logs remained from earlier stands (Carey and Peeler 1995). Thus, past selection harvest that removed the largest trees and snags probably did not favor northern flying squirrels, and, therefore, probably had negative effects on foraging by California spotted owls.

Blakesley (2002a, pers. comm. 2002) studied California spotted owls in an area where northern flying squirrels (
Glaucomys sabrinus
) comprised 70 to 80 percent of prey taken. She found that both survival and reproductive output were positively related to the proportion of the home range that was comprised of habitat selected by owls. Furthermore, site occupancy and reproductive output were negatively associated with the amount of non-habitat (non-forest and areas dominated by small trees and/or very low canopy cover).

In areas where the primary prey consists of dusky-footed woodrats, effects of timber harvest, either by selection or small patch cuts, may have been less severe on spotted owl prey. Dusky-footed woodrats are more abundant in shrubby areas than in areas with little shrub understory (Williams
et al.
1992), so this forage species may persist in harvested areas, at least at the lower elevations where it is more common. Franklin
et al.
(2000), who studied demographic performance of northern spotted owls in an area of northwestern California where dusky-footed woodrats were the primary prey, reported that adult owls with access to larger blocks of suitable forested habitat had slightly lower mortality rates, but those with home ranges that were more patchy with more openings had slightly higher fecundity (number of young produced per breeding female). A landscape pattern with some small patches of other habitats dispersed within and around a main patch of old forest appeared to provide the optimum balance in promoting both high fecundity and high survival. It seems likely that California spotted owls would have similar responses in the minority of their range where dusky-footed woodrats are the primary prey and thus may be less affected by habitat modification in those areas.

According to McKelvey and Johnston (1992), clear-cutting was the predominant harvest method on Sierran national forests only from 1983 through 1987. In areas where clear-cutting occurred during those years, and perhaps also where catastrophic fire has eliminated forested habitat, it may be reasonable to evaluate impacts based on studies of the effects of clear-cutting on the similar northern spotted owl. Bart (1995b) examined the relationship between amount of a northern spotted owl pair's home range that is suitable habitat and productivity and survivorship of owls. In Bart's (1995b) study area, habitat remaining after harvest was either of good quality (
i.e.
, remaining old growth) or very poor quality unsuitable for extensive use by owls (clear cuts). That analysis suggested that removing any suitable habitat within the vicinity of the nest tends to reduce productivity and survivorship of resident owls. Bart concluded that replacement rate reproduction might occur when 30 to 50 percent suitable habitat is retained within an owl's home range. However, he also noted that productivity and survivorship declined steadily below 80 percent suitable habitat and advised that northern spotted owl habitat should not be reduced to perceived thresholds in all instances or viability could be compromised. The primary form of habitat modification in the Bart (1995b) analysis area was clear-cutting. Therefore, these results may only have limited application to the California spotted owl, because much of the range of the California spotted owl has been selectively harvested. The selection harvest practiced in the Sierra is believed to have lowered habitat quality by removing large trees and snags, but it may not have rendered habitat completely unsuitable (USFS 2001a). Thus, the degree of impact of past selection harvest practices on California spotted owls remains unclear.

Spotted owl distribution in the Sierra Nevada is generally continuous and of uniform density within the historic range. However, several “areas of concern” were identified in Beck and Gould (1992). These are areas where densities of spotted owls are low, local populations are isolated, or distribution of habitat or owls is not continuous or is restricted because of past timber harvest, fire, and natural breaks in habitat. Areas of concern might be important if the range of the spotted owl begins to shrink. Beck and Gould (1992) identified 16 areas distributed throughout the range where there are gaps that delineate discontinuities in owl distribution (no habitat exists or there is a bottleneck) and 19 areas where concern relates to low population density, fragmented habitat, or loss of habitat due to fire.

The USFS (2001a) further cautioned that management in at least nine of these areas of concern in the Sierra Nevada could have disproportionate impacts to spotted owls without special management consideration. USFS noted that areas of concern that fall within checkerboard ownerships (blocks of private land interspersed with Federal lands) or fragmented habitats warranted special attention. Final management direction selected by the USFS-modified alternative 8 (USFS 2001b) included objectives for the amounts of habitat within each owl home range to provide for replacement rate reproduction.

Demographic Analysis

As one of the most intensively studied birds in the United States, the spotted owl has been the focus of research for well over two decades. Many sophisticated statistical techniques for estimating population trends have been developed and refined using data from the northern spotted owl, and the state of information for the California subspecies has benefitted accordingly. Across the range of the California spotted owl, five study areas (Lassen, Eldorado, Sierra, Sequoia-Kings Canyon, and San Bernardino), totaling about 2,200 square miles, have been established to examine the subspecies' population status. This research serves as a valuable resource for evaluating whether or not listing under the ESA may be warranted. In this section, we offer a synopsis and evaluation of the most current research on California spotted owl population trends. Because analytical techniques for assessing population status are complex, it is necessary to discuss the techniques, the studies, and their conclusions in some detail.

Several analytical methods have been applied to the analysis of population trend in spotted owls, and each method carries certain strengths and

weaknesses. Thus, to best understand population trend, it is important to concurrently assess the results of all methods instead of relying on a single analytical approach. One of the simpler methods uses raw empirical abundance data, where banded owls are counted and numbers are compared over time. Population trends can then be crudely assessed by evaluating abundance data from one year against similar data from a later year, or multiple years of data can be used in a regression analysis to determine the population trend from the slope of a regression line. While count data may appear straightforward, they are often subject to important sources of unquantifiable bias if the ability to detect owls changes from year to year. This can occur if survey effort changes over the course of the study or if the study area changes in size during the study period. Also, variation in detectability can be caused by environmental or behavioral factors. Numerous sources of possible bias can be present during the collection of abundance data in the field, especially over the long periods of time required to evaluate population trends in long-lived species such as spotted owls. However, basic abundance data can provide a reference point for comparison with the results of more sophisticated statistical methods, especially when possible error is reduced by careful data collection. Abundance data are available for each of the California spotted owl study areas, and are included in this evaluation by the Service.

Because of the problems that accompany abundance data, scientists have developed more sophisticated methods for estimating population trends that can be described in statistical terms, and which allow various statistical tests of the estimated population trend. These methods derive estimates of the annual rate of population change, otherwise known as lambda (λ), which is the fundamental measure for retrospective estimation of population trend. Varying analytical methods derive λ from data on vital rates (
i.e.
, birth and death rates) gathered using methods described for the northern spotted owl (Forsman 1983). Reproductive output is measured from direct observation of the number of young leaving the nest, and estimates of survival are obtained using mark-recapture techniques. Capture-recapture theory (Lebreton
et al.
1992) provides the foundation for deriving a statistical estimate of survival and population trend. In brief, this is done by capturing and uniquely marking individuals, and then recapturing (or resighting) those same individuals in subsequent years (Lebreton
et al.
1992). Some of the potential bias factors remain, such as variation in survey effort, but the recapture history for each marked individual serves as the basis for calculating vital rates for each age and gender class. After fecundity (
i.e.
, birth rate: number of female young fledged per female) and survival for the population are statistically estimated from field sampled data, those estimates are used to compute the finite rate of population change, or λ.

Lambda provides an estimate of two useful measures: the direction in population trend and the magnitude of population change (Franklin
et al.
1996). A λ value equal to 1.0 indicates a stationary population; less than 1.0 indicates a declining population; and greater than 1.0 indicates a growing population. The amount by which λ differs from 1.0 indicates the magnitude of the trend (
i.e.
, if λ = 1.10, the population has increased by an average of about 10 percent each year [1.10 to 1.0 = 0.10]). However, λ is a point estimate, and this estimate has a measure of precision. Therefore, researchers often test whether λ is significantly greater or less than 1.0, or equal to 1.0. For example, a λ = 0.97 may not be statistically different from 1.0 at some predetermined significance level if the confidence interval includes 1.0 (Lande 1988).

It should be noted that the estimate of lambda applies only to the period during which the data are collected. For this reason, long term studies are necessary to avoid misinterpretation of apparent trends. For instance, if a population demonstrates cycles that are completed over multiple decades, ten years of data may only capture a down cycle (which would falsely appear to be a decline) or up cycle (which would falsely appear to be an increase), depending on the timing of the study.

The five individual studies conducted on California spotted owl populations were consistent in their initial method for calculating λ, which has also been extensively described and applied in analyses of the northern spotted owl (Franklin
et al.
1996). In this method, survival and fecundity estimates for females were used in a mathematical tool called a projection matrix to solve for λ. Several issues may affect the validity of the projection matrix approach to calculating lambda. First, the method assumes that adult survival and fecundity are constant over time (Franklin
et al.
1996). Long term research on the northern spotted owl has demonstrated that this assumption is sometimes violated. Survival rates are not constant (Burnham
et al.
1996), and spotted owls have demonstrated variable annual fecundity, with occasional years of very high fecundity (Franklin
et al.
2002). However, the magnitude of the resulting bias appears to be small (Burnham
et al.
1996, Noon and Biles 1990). Second, individuals, particularly juveniles, may emigrate to areas outside the study area boundaries. Even though they could still be alive, these individuals are considered mortalities because they disappear from the study area, resulting in a survival rate that is biased low (Raphael 1996). To better understand the possible error in juvenile mortality rates, researchers compare the observed mortality rate with calculated theoretical rates that would be necessary for a stable population, and examine the difference. Although useful in some respects, this exercise does not alter the estimate of lambda for the subject owl population. The issue of juvenile emigration was addressed in the 1999 meta-analysis for the northern spotted owl, as well as for some of the individual northern spotted owl study areas, and overall trend estimates were adjusted for juvenile emigration (Franklin
et al.
1999).

Another potential issue regarding the projection matrix method is that the calculation includes only territorial birds (which are relatively easy to locate), ignoring nonterritorial, unlocated “floaters” that may be present and available to fill vacancies left by the eventual mortality of breeding birds (Franklin 1992). Bart (1995b) argued that the presence of floaters causes population trends to be determined by the trend in the amount of habitat, not by birth and survival rates. Using lambda estimates corrected for floaters and false juvenile mortalities, Bart (1995b) calculated that lambda estimates using the projection matrix method could be 0.13 to 0.03 lower than the actual value. Thus, for example, a population with an estimated lambda of 0.90 (signifying a decline of 10 percent per year) could actually be an increasing population. This argument should be considered in evaluation of lambda estimates. Trends in the nonterritorial segment of the population cannot be evaluated with the projection matrix method, although it is likely that over the long term, trends in the territorial and nonterritorial segments will follow similar trajectories, since they both depend on similar environmental conditions.

For these reasons, we approach the use of population matrix λ estimates with caution in this finding, and where possible, has sought additional

corroborative data and analyses before concluding that a population is declining. Our following discussion of the results from each of the five study areas will include evaluation of potential error in the lambda estimate that might result from these factors.

More recently, the data from the five study areas were reanalyzed using another statistical method. In 2001, owl researchers from the five California spotted owl study areas, timber industry consultants, and stakeholders met with experts in population analyses to conduct a meta-analysis of the available data (Franklin
et al.
2002). The term meta-analysis refers to the combined analysis of data collected from numerous studies to increase sample size and investigate relationships that would be difficult to assess with data from an individual study. A draft report authored by 15 participants (Franklin
et al.
2002) summarized the results of the five-day meta-analysis workshop and subsequent analysis.

It is our understanding that, as of the publication of this finding, peer review comments have been received by the authors of the meta-analysis, but the incorporation of peer review comments by the authors has not been completed. Thus, the meta-analysis manuscript remains a draft. We have examined the draft meta-analysis document, the comments of prominent peer reviewers, and solicited comments from the authors regarding our conclusions herein. We regard the draft meta-analysis as the best available science on the subject, but as stated above, we have not relied solely on this analysis in developing our conclusions regarding population trend.

The meta-analysis of adult survival was based on female and male adult capture histories for the five study areas, but fecundity was estimated for each study area separately because differences existed in field sampling protocols. To eliminate a possible bias in projection matrix estimates of λ due to inaccurate rates of survival (resulting from unknown emigration rates), a new technique was used to calculate λ, called the “temporal symmetry capture-recapture model” (Pradel 1996). Pradel's method calculates the rate of change in population size between two successive years using mark-recapture histories for each owl, and since this technique calculates annual estimates, λ can change each year. In contrast, the projection matrix method calculates an average λ estimate for the period of study using a population's average birth and death rates. Pradel's measure applies to subadult and adult territorial owls, and incorporates birth, death, emigration, and immigration rates. Estimates of juvenile survival are unnecessary because movement of spotted owls into and out of the study area is considered in changes of owl numbers over time.

While Pradel's λ accounts for permanent emigration of juveniles, it doesn't provide insight as to the root cause of a population's rate of change. For example, if λ = 1.0, indicating a stable population, it is impossible to know if the stability is a result of immigration or new recruits from births, which prevents inferences about the health of the local population (Franklin
et al.
2002). Thus, it is important that trends in survival and fecundity rates be examined concurrently with assessments of λ. Pradel's λ provides information as to whether owls are being replaced from within or outside the study area, and not solely whether they are replacing themselves, which is the goal of the projection matrix approach. Because the Pradel method provides an estimate of one λ for each year, the annual λ estimates can themselves be assessed for trends, and a mean estimate can be calculated for the period of study.

Franklin
et al.
(2002) applied the Pradel method to each of the five individual study areas, and conducted a combined meta-analysis of the results from the four study areas (Lassen, Eldorado, Sierra, and Sequoia / Kings Canyon) that lie in the Sierra Nevada. The following discussion details the results of the earlier projection matrix analyses and reports of basic count data for each study area, and compares those results with the new results derived using the Pradel method, as reported by the draft. Following the discussion of individual areas, we will describe the results of the meta-analysis of the four combined Sierra Nevada study areas.

Since survey areas changed throughout the course of some studies, only those areas (within larger study areas) that received surveys from start to finish were included in the new analysis, and only years that received consistent survey effort were used in the analysis (Lassen study area [490 mi
2
]: 1992 to 2000; Eldorado study area [137 mi
2
]: 1990 to 2000; Sierra study area [137 mi
2
]: 1990 to 2000 Sequoia and Kings Canyon study area [132 mi
2
]: 1991 to 2000; San Bernardino study area [730 mi
2
]: 1991 to 1998).

In this review, we primarily used the most recent report or published article for each area, although we reference earlier reports to clarify apparent changes in results for a given area. We summarize the results below.

Lassen Study Area—The Lassen study area encompassed approximately 850 mi
2
in northeastern California, the majority of which was located in the Lassen National Forest. Small segments of the study area included the Plumas National Forest, private timber lands, Lassen Volcanic National Park, and Bureau of Land Management land. According to Blakesley and Noon (
in litt.
2003), four lines of evidence suggest that the Lassen population has been decreasing. However, this information, received very recently, could not be fully evaluated. The most recent publication for this study area (Blakesley
et al.
2001) covered 10 years of field sampling (1990 to 1999), during which the annual rate of population change was estimated to be 0.910 using the projection matrix method. This estimate was significantly less than that of a stationary population (λ = 1.0), and suggested that the territorial female owl population (those females that occupy and defend a habitat area) declined 9 percent annually from 1990 to 1999. Blakesley
et al.
(2001) inferred that if the conditions present during their study remained constant into the future, and if the true rate of change were as low as 4 percent instead of the estimated 9 percent, that the population would decline by one-half within 20 years. Such forecasting beyond the period of data collection is unreliable, and the accuracy of this projection is likely biased, as conditions are unlikely to remain constant for 20 years (Burnham
et al.
1996, Raphael
et al.
1996, Noon
et al.
1992).

We were unable to compare the estimated value of λ to the observed numbers of territorial adults in this study, because the survey area increased over time. However, from the estimated growth rate of 0.910, we can conclude that over 50 percent of the population would be lost by the end of the study. According to Blakesley and Noon (
in litt.
2003), within 68 territories surveyed consistently from 1993 to 2001, the number of female owls declined from 56 to 37. This suggests a decline of 5 percent annually, which is not statistically different from the 9 percent decline estimated above. A potentially large source of error arises from unknown rates of juvenile and adult emigration. Blakesley
et al.
(2001) suggested that while incorrect juvenile emigration rates may have resulted in a survival estimate biased low, the magnitude of the bias was probably small. For the Lassen population to demonstrate a stationary trend during the study period (given that all other vital rates were accurate, including an

adult survival probability of 0.827), the juvenile survival rate would have to more than double (from the estimated 0.333) to 0.790. However, given that all other parameters remain the same, Franklin (2003) estimated that if adult and subadult survival was actually 0.85, juvenile survival would have to be 0.657 to achieve a stationary population, and if adult and subadult survival were actually 0.87, juvenile survival would have to be 0.55 for a stationary population. A juvenile survival probability of 0.55 is within the realm of possibility based on juvenile survival estimates for northern spotted owls on two study areas (Franklin
et al.
1999).

The meta-analysis for this study area included nine years of sampling (1992 to 2000) and encompassed 490 mi
2
. The mean λ estimate for the period using Pradel's method was less than 1.0 (0.985), but was not statistically different from that of a stationary population (λ = 1.0). Examination of the annual λ estimates (per year, as opposed to the above mean λ) showed no evidence of a trend for the Lassen study, and adult apparent survival showed no substantive variation or trends through time. Fecundity was so variable through time that a linear trend (as opposed to sporadic high-low trends) could not be identified.

Although there is information that suggests that this population may be declining, uncertainties exist when interpreting the projection matrix approach, actual counts, and Pradel's methodology. Without further refinement of the projection matrix approach (
i.e.
, adjusting juvenile survival estimates using radio telemetry), it is difficult to reconcile the declining projection matrix λ of 0.910 with the statistically stationary estimate of 0.985 derived using Pradel's methodology. Thus, we cannot conclude with certainty that the population is declining, increasing, or stationary.

Eldorado Study Area—The Eldorado study area consisted of two segments: a 137-mi
2
density study area, and a 220-mi
2
regional study area. The most recent publication for this study area (Seamans
et al.
2001a) covered 10 years of field sampling (1990 to 1999). Although surveys took place from 1986 to 1999 in the density study area, surveys in the regional study area were initiated in 1997. Only data from 1990 to 1999 were used because survey effort and sample sizes increased dramatically after 1989 due to increased funding (Seamans
et al.
2001a). The study area was located primarily within the Eldorado National Forest, but portions were also located within the Tahoe National Forest and the Tahoe Basin Management Area. Researchers lacked sufficient data to calculate the juvenile survival rate on the Eldorado study area, so they used the survival rate from the nearby Lassen study area (0.333) as a surrogate. This estimate was thought to be optimistic, as estimates of northern spotted owl juvenile survival from 11 study areas averaged 0.258 (Forsman 1996), and in the Eldorado study area, 11 of 147 individuals banded as juveniles were recaptured as territory holders, which would translate to a survival probability of 0.074 (Gutie

rrez
et al.
2001). However, there is a likelihood that the estimated juvenile survival of 0.258 for northern spotted owls was underestimated, as it was not corrected for juvenile emigration. A later report (Franklin
et al.
1999) adjusted juvenile survival estimates in three northern spotted owl study areas to reflect juvenile emigration rates calculated from radiotelemetry data. The adjusted juvenile survival rates were 0.598, 0.632, and 0.366. These estimates represented increases of 137.2%, 41.8%, and 87.9% in juvenile survival estimates for each respective study area.

Using the projection matrix approach, the annual rate of population change was estimated to be 0.948, which was significantly less than that of a stationary population. This λ value suggested that the territorial female owl population declined 5.2 percent annually from 1990 to 1999. In contrast, female abundance at the start (1990) and finish (1999) of the study was 26 and 28 individuals, respectively. This difference between the estimated λ and the rate calculated from actual numbers could be attributed to immigration of individuals into the study area. If true, this would indicate that individuals were not replacing themselves, but were being replaced by recruits from outside the study population. Earlier estimates of λ from this study area calculated similar trends (λ= 0.947) using only 6 years of data (Noon
et al.
1992). This six year estimate was not statistically less than 1.0, but the power (ability to detect differences) of this test was low, so the trend of the population was uncertain at the time (Verner
et al.
1992). Results from the 2001 study (Seamans
et al.
2001a) expanded the sample size and study period, and increased the statistical power of their test so that their estimate of lambda (0.948) was then found to be statistically less than 1.0.

The use of a surrogate juvenile survival rate in this study may introduce bias into the estimate of λ for two reasons: the Lassen estimate of juvenile survival probably carries certain biases given the inability to consider juvenile emigration in the estimate of juvenile survival, and the Lassen study area may not accurately represent the Eldorado study area. Further, Gutiérrez
et al.
(2001) reports that survey effort for this study area changed over time, and that survey effort can influence density and survival estimates. For this reason, data from the first four years of study were not included in the estimate of survival or reproduction (Seamans
et al.
2001a analyzed data starting in 1990, not 1986). In a subsequent report (Seamans
et al.
2001b), the projection matrix estimate of λ was compared to a growth rate calculated from actual numbers of adult females present during the study. The growth estimate from actual numbers was 0.951, and was significantly less than zero. This estimate was calculated using data from 1993 (37 adult females) to 2000 (24 adult females), while the value derived from the projection matrix approach (0.948), which was calculated using data from 1990 to 1999.

The meta-analysis for this study area included 11 years of sampling (1990 to 2000) and encompassed 137 mi
2
. The mean λ estimate for this study area was greater than 1.0 (1.042), and was not statistically different from that of a stationary population (λ = 1.0). Examination of annual λ estimates showed a significant decline, and similar to the Lassen study area, adult apparent survival showed no substantive variation or trends through time. No linear trend in fecundity could be identified.

While the projection matrix estimate of λ showed a decline, (and trend in Pradel's annual λ showed a decline), actual counts increased a small amount from 1990 (26 adult females) to 1999 (28 adult females) but decreased from 1993 (37 adult females) to 2000 (24 adult females), and the mean λ estimate using Pradel's method appeared to show a stationary population. Furthermore, there were substantive uncertainties regarding the accuracy of vital rate estimates used in the projection matrix estimate of λ. These results do not allow us to reach a definitive decision with respect to population trend on the Eldorado study area, and we cannot conclude the population is declining.

Sierra Study Area—The Sierra study area was located primarily (92 percent) within the Sierra National Forest, and encompassed the watersheds of the San Joaquin River and the North Fork of the Kings River (Franklin
et al.
2002). The study area included approximately 263 mi
2
, and the boundaries were delineated based on National Forest boundaries and major topographic features such as

ridges and drainages. Spotted owl telemetry studies and intensive surveys began in 1987 and 1990, respectively, on a 160 mi
2
portion of this study area (old Sierra study area). In 1994, surveys were expanded to include an additional 103 mi
2
(new Sierra study area; Steger
et al.
1999). Juvenile survival rate was not calculated using data from this study area. Instead, the juvenile survival rate from the San Bernardino study area (0.328) was used to approximate the Sierra study area's juvenile survival rate.

Using survey data from 1990 to 2000 and the projection matrix method, the annual rate of population change was estimated for the old Sierra study area (1987 to 2000) and both old and new Sierra study areas combined (1987 to 2000). Annual rates of population change for the old Sierra and combined Sierra study areas were 0.897 and 0.901, respectively. These estimates were significantly less than that of a stationary population, and suggested that the territorial female owl population declined about 10 percent annually from 1987 to 2000. For an 11 year period (1990 to 2000), this translates to a population decline of around 60 percent. For the old Sierra study area during 1991 to 2000 (1990 was not examined as survey guidelines were not yet established on the study area), actual owl numbers seemed to corroborate a decline, albeit the drop in numbers was less severe than 60 percent. Owl abundance in 1991 and 2000 were 69 and 55, respectively. These numbers represent a 20 percent decrease, although the accuracy of the count numbers is unknown. The new Sierra study area also showed a decline: actual owl numbers dropped from 37 in 1994 to 29 in 2000.

The meta-analysis for this study area included 11 years of sampling (1990 to 2000) and encompassed 137 mi
2
. The mean λ estimate for this study area was less than 1.0 (0.961), but was not statistically different from that of a stationary population (λ = 1.0). Annual λ estimates showed a weak (nonsignificant) decline, and adult apparent survival showed no substantive variation or trends through time. The Sierra study showed a negative trend in fecundity, which could have been driven by a high reproduction year early in the study.

Although the mean λ was statistically stationary using Pradel's methodology, actual numbers of owls declined; the projection matrix approach showed a decline; there was a negative trend in fecundity; and there was a weak, nonsignificant decline in annual λ estimates (using Pradel's method). It appears that a decline in this study area is possible, but the use of a surrogate juvenile survival rate introduced a bias of unknown proportion, and thus, the magnitude of a possible decline remains uncertain.

Sequoia and Kings Canyon Study Area—The Sequoia and Kings Canyon study area encompassed approximately 130 mi
2
of land in Fresno and Tulare counties. The majority of the area was located in the Sequoia and Kings Canyon National Parks. Small segments of the study area include the Kings River watershed, but most of the study area was in the Kaweah River watershed (Franklin
et al.
2002). Surveys in this study area cover 11 years of field sampling (1990 to 2000), but useful data exist from a previous demographic study that began in 1988 (Steger
et al.
2000). Demographic surveys were conducted on130 mi
2
of land in this area, and methods for calculating λ were identical to those used for the Sierra study area. The annual rate of population growth using the projection matrix method was estimated to be 0.973, suggesting a decline of 2.7 percent per year. Statistical testing found that λ was not significantly less than 1.0. Actual owl counts during the study period seemed to indicate that the population might be growing, but again, the accuracy of such numbers is uncertain. Owl abundance in 1990 and 2000 were 54 and 64, respectively.

The meta-analysis for the Sequoia and Kings Canyon study area covered 132 mi
2
during 1991 to 2000. The resulting mean λ estimate was 0.984, but was not statistically different from that of a stationary population. A significant quadratic trend (decline, then increase) was detected for annual λ estimates, but adult apparent survival showed no substantive variation or trends through time. A linear trend could not be identified in fecundity estimates, as fecundity was highly variable through time. Apparent survival for the Sequoia and Kings Canyon study area was significantly higher (0.877) than that of the other study areas combined (0.819).

Lambda estimates using the projection matrix approach and Pradel's method suggest stationary population trends, and actual owl numbers do not show declines. Trend in annual λ estimates also does not show a decline, and apparent survival in this study area was higher than all other study areas examined in this finding. Based on these results, and considering the inclusion of the juvenile survival rate from another study area, we cannot conclude that this population is declining.

San Bernardino Study Area—The San Bernardino study area was located entirely within the San Bernardino National Forest, and comprised all suitable habitat for spotted owls within the mountains. Surveys covering approximately 200 mi
2
(Big Bear study area) began in 1987, but were expanded in 1989 to cover the entire San Bernardino Mountain range (approximately 730 mi
2
; San Bernardino Mountains study area; Gutiérrez et al 1999). This study area was unique in that it exists in southern California as a relatively isolated population (Gutiérrez and Pritchard 1990, LaHaye
et al.
1994). Higher elevations in the study area contained forested habitat suitable for spotted owls, while lowland areas of unsuitable desert scrub and chaparral habitats surrounded and isolated the higher peaks (Noon and McKelvey 1992). Early projection matrix studies using four, five, and six years of data estimated significant annual declines during years between 1987 and 1993 (λ = 0.769, 0.827 and 0.860, respectively; LaHaye
et al.
1992, Noon
et al.
1992, LaHaye
et al.
1994). The annual rate of population change for the most recent report we possess was estimated to be 0.91 based on 11 years of data (1988 to 1998; LaHaye
et al.
1999). This estimate was significantly less than that of a stationary population, and suggested that the territorial female owl population declined nine percent annually from 1988 to 1998. Over the 11-year study period, this rate of decline would translate to a loss of over 60 percent of the population.

Although all forested habitat within the San Bernardino Mountains study area (including unoccupied habitat) was surveyed (Gutiérrez 2001), survey effort increased during the study period. In 1989, 532 total surveys were conducted, whereas in 1998, 1,185 total surveys were conducted (LaHaye
et al.
1999). This change in survey effort could cause the number of owls observed in any year to be a function of the survey effort instead of an actual trend in numbers. Thus, there is a high likelihood that an assessment of actual owl numbers through time could be biased, and may not accurately represent the true population size in any given year.

The meta-analysis for this study area included 8 years of sampling (1991 to 1998) and encompassed 730 mi
2
. The mean λ estimate for this study area was less than 1.0 (0.978), but was not statistically different from that of a stationary population (λ = 1.0). Examination of annual λ estimates showed a weak (nonsignificant) decline, and as with all other study areas, adult apparent survival showed no

substantive variation or trends through time. No linear trend in fecundity could be identified.

Although the projection matrix approach showed a decline and there was a weak, nonsignificant decline in annual λ estimates using Pradel's method, the mean λ was statistically stationary using Pradel's methodology. This is more meaningful than with the other study areas. Recall that using Pradel's method, the population could appear stationary or growing, even if the population growth is caused by outside immigration. For this study area, recruitment cannot likely be attributed to immigrants entering the study area, as the study area is relatively isolated. Thus, the disparity between the projection matrix estimate and the estimate using Pradel's method could be a result of either (1) measurement error in survival rates of juveniles or adults, or (2) the presence of unlocated floaters in the study area (Franklin 1992). Regardless, we have insufficient certainty as to the status of this population to conclude that it is either declining, increasing, or stationary.

Meta-Analysis Results

The meta-analysis used the Pradel method to evaluate data from various study areas, as described above. The analysis was also applied to some aspects of the population as a whole. The estimated mean lambda for each of the individual studies was not significantly different than 1.0 over the periods analyzed. The meta-analysis did not estimate a mean overall value of lambda for all study areas combined. Examination of trends in annual λ over the periods analyzed for each study area showed no evidence of a trend for the Lassen study, a significant decline for the Eldorado study, a significant quadratic trend (decline, then increase) for the Sequoia and Kings Canyon study, and a weak (nonsignificant) decline for the Sierra study. The overall trend in annual lambda rates for these four Sierra Nevada study areas was declining, then increasing. The trend in annual rates for the San Bernardino study was a weak (nonsignificant) decline for the period analyzed.

Adult survival rate for the Lassen, San Bernardino, Sierra, and Eldorado study areas combined was 0.819, which was substantially lower than the mean estimate (0.850) for adult northern spotted owls across 15 study areas (Franklin
et al.
1999). Except for the Sierra study area, fecundity estimates for the California spotted owl study areas were so variable through time that linear trends could not be identified. The Sierra study showed a negative trend in fecundity, which could have been driven by a high reproduction year early in the study.

Conclusions—In total, the findings reported above are not conclusive with respect to the population status of the California spotted owl. There is no definitive evidence that the population is decreasing across its range, and various analytical results of the individual study areas are not wholly supportive of conclusions regarding declines in any given study area. Low levels of declines may be occurring in some study areas, but if so, they are not clearly evident using existing analytical techniques. The strongest support for a possible decline is on the Sierra study area, and the strongest support for a possible stationary population is on the Sequoia and Kings Canyon study area. The combined rate of adult survival for all study areas except the Sequoia and Kings Canyon study area may be of concern, as it is substantially lower than that reported for the northern spotted owl. However, Pradel's λ estimate for each study area consistently showed statistically stationary populations, so we cannot conclude that this lower adult survival rate is causing a decline in California spotted owl populations. At this time we have no clear statistical evidence to show that the California spotted owl is declining throughout its range.

Summary of Factors Affecting the Species

Section 4 of the Act (16 U.S.C. 1531
et seq.
) and the regulations (50 CFR part 424) that implement the listing provisions of the Act set forth the procedures for adding species to the Federal lists. A species may be determined to be an endangered or threatened species due to one or more of the five factors described in section 4(a)(1). These factors and their application to the California spotted owl are as follows:

A. The Present or Threatened Destruction, Modification, or Curtailment of Its Habitat or Range

There are two categories of possible threats to the California spotted owl that are related to habitat. The first is the current threat related to the condition of existing populations and habitat, and existing, ongoing habitat modification. The second is the potential threat that may result from future management of habitat.

Threats related to current condition of populations and habitat.
Numerous authors have expressed concern over the current status of the California spotted owl, and much of this concern is related to the quality of the habitat available to the subspecies at the present time. The best scientific information available indicates that high survival of spotted owls is achieved by maintaining large, unfragmented areas of suitable habitat (Moen and Gutiérrez 1997, Franklin
et al.
2000, Blakesley
in litt.
, 2002a). Important habitat components, especially large trees, large snags, and large down logs, are currently in short supply across the range of the California spotted owl (Verner
et al.
1992b, USFS 2001a). The diameter of nest trees selected by owls in the Sierra Nevada is significantly greater than the average diameters of conifers in the Sierra Nevada. Large trees become future large snags and large downed logs, the latter providing important habitat attributes for some prey species. The length of time required to recover old trees and increase their density over the landscape raises the level of concern associated with their decline (USFS 2001a).

Concern has been exacerbated because, although harvest volume was declining markedly in the 1990s, existing management direction did not appear to be sufficient to arrest completely habitat decline (USFS 2001a), and because, until recently, most of the remaining old growth in national forests in the Sierra Nevada remained in areas available for timber harvest (Franklin and Fites-Kaufmann 1996, Beardsley
et al.
1999). Thus, Blakesley and Noon (1999) argued that the most positive step that can be taken to reverse apparent declines of California spotted owls would be to increase retention and recruitment of large trees and closed canopy conditions throughout the Sierra Nevada.

Concern also exists because the existing habitat used by California spotted owls appears to be vulnerable to stand-replacing catastrophic fire. Removal of large overstory trees in conjunction with fire exclusion has lead to changes in forest structure that favor spread of high intensity fire. Conifer stands have become denser and composed mainly of trees in small and medium size classes (University of California 1996). The species composition of these forests has also shifted, from more shade intolerant, fire-hardy species such as ponderosa pine and black oaks to more shade tolerant, fire sensitive species such as white fir and incense-cedar (Verner
et al.
1992, Weatherspoon
et al.
1992). Similar increases in density and changes in species composition have been documented for coniferous forests of southern California (Weatherspoon
et al.
1992, Minnich
et al.
1995). Forests in

southern California were logged significantly into the 1960s (Verner
et al.
1992a) and have also been affected by intensive fire suppression. Dense stand conditions in California forests have lead to increased tree mortality, due to competition, drought, insects, disease, and, in some cases, air pollution (University of California 1996). The increased density and continuity of young trees together with increased fuels from fire suppression and tree mortality have created conditions favorable to more intense and severe fires (University of California 1996). Over the last 30 years, the annual acreage burned in the Sierra has been about 19,020 ha (47,000 ac), but in the last ten years, that average was 30,756 ha (76,000 ac) (USFS 2001b).

Paradoxically, the growth of understory trees that contribute to the high degree of canopy closure favored by California spotted owls also increases the risk that wildfire might spread into the canopy and destroy the stand. Because fire suppression has increased density of stands in the Sierra Nevada, the possibility exists that it has led to net improvement in owl habitat in some areas (Weatherspoon
et al.
1992) with resultant increases in spotted owls (Verner
et al.
1992a). Weatherspoon
et al.
(1992) characterized forests selected by spotted owls as having the structural components favorable for crown fires.

However, in recent years, relatively few California spotted owl sites have been severely impacted by wildfire (USFS 2001a). From 1993 to 1998, only 15 Protected Activity Centers, a 121 ha (300 ac) management area established around all Forest Service owl sites in 1993, burned in wildfires, and three of those remain occupied (USFS 2001a). This possibly is because of the success of initial attacks on wildfires in Sierran mixed conifer types (Weatherspoon
et al.
1992) or because California spotted owls often occupy relatively moist areas such as northern aspects (Gould 1977, Barrows 1981, Gutiérrez
et al.
1992, North
et al.
2000), lower slopes of canyons (Gould 1977, Gutiérrez
et al.
1992), or areas close to water (Gould 1977) where average fire intervals are longer (Weatherspoon
et al.
1992). Weatherspoon
et al.
(1992) noted that most of the large fires in the Sierra Nevada had occurred on the eastside and at low elevations on the westside, outside the areas where spotted owls occur.

During development of the SNFPA, the overall concerns r

[Text truncated at 120,000 characters. The full text is on the page linked above.]

---

Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/fr%3A03-3519. Public record. Not legal advice.