Methods - Wildlife Species
This section of the Methods addresses wildlife; it is organized as follows:
Determining Wildlife Species Information
Determining the Wildlife-Habitat Information
Natural History of Selected Species
Basin-wide Wildlife Habitat Type Maps
Key Ecological Functions of Fish and Wildlife Species
To analyze wildlife at the broad scale of the entire Columbia River Basin in the U.S., we used “coarse-grain” information on environmental conditions. Analyses assess data and information on occurrence and change in overall wildlife habitat types for each 6th field HUC (subwatershed), summarized to larger areas of provinces and to the Columbia Basin as a whole; this is what constitutes “coarse-grain information.” Results of these wildlife analyses should be viewed most appropriately at scales of provinces and the entire basin, as defined in the Fish and Wildlife Program Scientific Foundation (see Appendix L). Wildlife analyses herein provide broad distributional patterns of habitats, potential species occurrence, and ecological functions.
Subsequent stages of the Framework analysis will summarize
results from the 6-HUC to the subbasin level and then to the province and
basin. Accuracy at levels finer
than the province level will entail using finer-grain information on the
distribution and changes in structural conditions and specific substrates and
other influential conditions (key environmental correlates) within each
wildlife habitat type and within each subwatershed. We did not have that
information available for the current analyses.
Our analysis provides three “snapshots” of historic,
current, and future habitat conditions. Historic
conditions (in places referred to as Historic Potential) refers to conditions
in the absence of non-native human influence, projecting back in time,
approximating conditions that may have generally occurred during the early 19th
century. Maps of historic
conditions should be viewed as general patterns and not as depicting specific
conditions within watersheds. Current
condition (in places also referred to as Current Potential) is estimated using
data on vegetation, which we interpreted as wildlife habitat, collected within
the last ten years (i.e., 1995-2000). Data on current vegetation conditions,
in most cases, has not undergone formal validation but is based on spectral
classification of satellite imagery that has been related to land use/land
cover types, that in turn have been determined from point or field locations.
Future conditions refers to wildlife habitats that would stabilize over
the long term, say on the order of 50-100 years, under the management
strategies specified under each of the three planning alternatives we
analyzed. L. Vail (Appendix L)
developed a screening procedure to estimate the acres of wildlife habitat
types for seven alternatives, three of which are used in this analysis. This
procedure estimated the degree that current wildlife habitat types would shift
back towards the historic wildlife habitat types for a specific alternative.
Each alternative was composed of a set of strategies.
The effectiveness and intensity of each strategy, as defined by
McConnaha, et al. (2000), were considered in combination with the land use and
land ownership for each 6-HUC. Spatial information was not considered at a
spatial scale finer than the 6-HUC. This approach was consistent with the
other elements of the Multi-Species Framework Project. It implies that no
information about adjacency of habitat types and land ownership is known. For
instance, we don’t know if forest represents riparian or non-riparian
regions. In our assessment, there
is one historic condition, one current condition, and three alternative future
conditions (one future condition for each of the three alternatives analyzed
in this report).
Comparing historic to current conditions provides some
understanding of how habitats have changed from recent land use practices.
This provides a baseline by which to compare current to future
conditions under each of the three planning alternatives we analyzed.
Future potential conditions represent a range of possible future
changes and whether alternatives are moving wildlife habitat quality and
quantity toward or away from historic habitat conditions. This process of comparing alternative performance to current
and historic performance for wildlife is similar in concept to the fish
analysis process but several details differ:
The fish analysis uses historic, current and future conditions
to set biological objectives for specific fish habitat attributes (usually
fine-scale substrates, which we refer to as key environmental correlates, or
KECs). Because of the broad scale
addressed, the wildlife analysis is not based on key environmental correlates
but rather on more general wildlife habitat capacity summarized at the scale
of the entire basin. Since the proposed terrestrial wildlife habitat changes are
relatively small in area compared to the area of the whole basin, any change
in the percentage of wildlife habitats (that is, conversion of one wildlife
habitat type to another type) is very small (e.g., 0.01 percent), and not
particularly useful for setting biological objectives at a basin scale.
When fine-scale KEC data are available within subbasins, it will be
more appropriate to set local biological objectives for wildlife.
The fish EDT method assesses three demographic components of
fish species performance: survival, capacity and life history diversity.
These components can be combined and expressed in terms of fish
population density for any of the historical, current or future conditions.
Wildlife performance is expressed as habitat capacity. Survival, life
history diversity and density of wildlife are not explicitly assessed at the
basin level.
Fish performance expressed as density was estimated using a
Beverton-Holt model, which includes both density-independent and
density-dependent components. The density-dependent component of EDT is based
on habitat quality times area. When
habitat quality is less than optimal due to one or more management activities
(i.e., mining), the reduction of habitat quality is used to reduce population
density estimates. As such,
optimal habitat conditions for a particular area are reduced as a function of
the type, the intensity and the effectiveness of management activities.
The actual population density response may or may not meet the model
projections due to a variety of factors that are not related to habitat (e.g.,
hunting and toxics). Hence
modeled conditions are referred to as potential: historic potential, current
potential, and future potential. Wildlife
performance, based on habitat capacity, does not include an estimator of
density but does include an estimator of habitat quality based on the
intensity and effectiveness of a management activity (i.e., collection of
strategies) and the resulting influence on the percentage of wildlife habitat
types (Vail, 2000, pers. comm.).
Analyses of species’ key ecological functions are based on
the Northwest Habitat Institute’s Species Habitat Project (SHP) database.
The SHP database was initially developed for Oregon and Washington and
built upon in part from the efforts of the Interior Columbia Basin Ecosystem
Management Project (ICBEMP) to include the entire Columbia Basin in the United
States. Increasing the area of
coverage for the SHP database to the whole basin within the United States
resulted in adding 12 wildlife species to the SHP database. We have made progress in including fish key ecological
functions into this expanded database, but this work is not complete.
One of the fundamental features of the EDT approach is the relationship between two dimensions: habitat attributes and population productivity across the landscape. Population productivity, a dimension that drives the Ricker and Beverton-Holt models, is important to fishery biologists for setting harvest management guidelines. The implied relationship between habitat, productivity, and fish management seems reasonable but recent calls for an ecosystem–based approach imply that an additional dimension of processes, beyond habitat description, may be at play. For example, Rose (2000) cited community-level interactions, habitat complexities, and cumulative effects as three of six issues that have prevented fisheries managers from achieving their goal of sustainability. Cederholm et al. (2000), in a manner similar to Rose, discussed the need for an ecosystem approach to understand the cumulative impacts of human development on fisheries management. These, as well as other authors, call for new tools to assess these processes (community interaction, habitat complexities and cumulative effects) and to provide an additional dimension of insight to the relationship between habitat and regaining lost fish productivity and diversity.
Schlosser and Kallemeyn (2000), on the heels of the above articles, opined that there has been a fundamental shift in ecology toward a broader geographic perspective that incorporates hierarchically-structured and scale-dependent levels of variation and complexity. Their insightful article on the spatial and temporal relationship between beaver and fish abundance and diversity not only clearly demonstrates the importance and roles of wildlife-generated structures in the aquatic environment, but also documents the importance of the additional dimension of ecosystem functions as beaver dams are built, abandoned, collapse, and rebuilt in relation to changing hydrologies in space and time. The documented relationships between higher fish species diversity and the collapsed successional beaver ponds, and between higher fish abundance (lower diversity) with intact beaver dams and the nonrandom distribution of successional environments on the landscape, indicate that we need to consider the third dimension of habitat analyses, species, and ecological functions, across landscapes, to better understand and predict relationships between habitat and fish abundance and diversity.
The Framework proposes to assess this third dimension of habitat using the concept of key ecological functions (KEF) developed by Marcot et al. (1997) and expanded by Marcot and Vander Heyden (2000). Analyses of KEFs across the landscape in the Framework hierarchical process can provide insight to both upland and aquatic functional webs that may, as Rose (2000) opined and Schlosser and Kallemeyen (2000) demonstrated, improve the understanding of how habitat attributes influence fish and wildlife sustainability.
A variety of concepts (e.g., energy flow, trophic relationships, indices of biological productivity, Lotka and Volterra predator and prey equations, and aggregated functional groups such as guilds) are available to investigate fish and wildlife interaction (Powers et al. 1995, Karr 1991, Rose 2000, and deMaynadier and Hunter 1997). These and other concepts were reviewed and considered by the Ecological Work Group. While the many concepts available to us offered their own unique insights into community relationships, none of them offered the opportunity to quickly assess and integrate the ecological functions of all species in the community from a common database. Furthermore, none of the concepts available to us could be readily linked to the population models (e.g., Beverton-Holt and Ricker) that are important to fisheries managers.
The EDT approach, which uses a Beverton-Holt hierarchical-landscape model, offers several opportunities for community level input as Level 2 Attributes. For example, beaver ponds, predation, and community interactions are Level 2 Attributes that relate directly to wildlife functions. Thus our ability to quantify and rank these Level 2 Attributes for the same landscape units (i.e., HUC 6) used by EDT will improve the quality of the variables that drive the EDT landscape assessment.
We used two basic approaches to quantify wildlife species and community input to EDT. Both of these approaches are closely associated with and rely on the SHP Wildlife Habitat Relationship (WHR) database (Johnson and O’Neil 2000). The first approach used wildlife habitat-capacity models to assess the likelihood that specific wildlife functions, such as beaver dams, are likely to occur in a given HUC 6. A Habitat Condition Index (a HEP-type analysis for large landscapes) has been developed to assess habitat condition for individual species. Output from this methodology can be the basis for: (1) ranking EDT Level 2 Attributes such as the presence of beaver ponds and (2) assigning a weight to a particular species to assign a proportional contribution to a particular ecological function. For example, a predatory species in marginal habitat (low HCI) is not likely to be a large contributor to the piscivory function.
The second approach is to harness key ecological function (KEF) analyses (Figure III.B.1) to assess EDT Level 2 Attributes such as Community Richness. For example, functional redundancy is calculated for all wildlife species, performing a function per the WHR database, for each HUC6. While this analysis is currently limited to wildlife, there is ample opportunity for including fish and wildlife functions in this assessment method. Fish functions for chinook and bull trout have been successfully integrated with wildlife function in pilot analyses. Functional redundancy is but one of several functional analyses that can be quantified in a hierarchical analysis to examine patterns of the “functional web” of fish and wildlife.
The wildlife analyses in this report, at the basin level, are based on wildlife habitats assessed from coarse digital data from a variety of databases. They are intended to be instructive and focus on the Framework vision of a multispecies process with a salmon emphasis. We illustrate two ways fish and wildlife can be integrated and offer a procedure to assess functional influences (for fish and wildlife) that can stand on its own as well as be related to salmonid models (i.e., EDT). At the basin level we will use our analyses to question whether alternatives/strategies for salmonid recovery will: (1) provide similar benefits for wildlife, (2) have a negative influence on wildlife and (3) potentially influence functional linkages between fish and wildlife (and visa versa). This initial basin analysis is only the beginning in that it provides a Framework for similar analyses at the province and subbasin levels of the landscape hierarchy. As similar types of analyses are conducted at the lower hierarchical levels, with a finer scale of data and area specific models, the lower level analyses should be re-aggregated (Stienetz et al., 1999) at the basin level to re-assess the basic interactive questions cited above.
To summarize, the Framework process has developed a common
platform (database, methodology and theory) for assessing fish and wildlife
populations and ecosystem function in the Columbia Basin.
Several applied examples are presented in the following report and
suggestions for database and methodological advancements are presented.
In addition, the platform provides the foundation for addressing
ecosystem goals and biological objectives at various hierarchical scales
within the basin.
The wildlife information that supports the Framework analysis comes from a 4-year project that recently updates databases on wildlife-habitat relationships in Oregon and Washington (Johnson and O’Neil 2001). We refer to this project as the Species Habitat Project (SHP) that was initially developed for Oregon-Washington and later expanded to include the U.S. portion of the Columbia River Basin. The SHP data set has information on 593 wildlife species occurring in Oregon and Washington; this data set was modified to include an additional 12 wildlife species known to occur in Idaho, Western Montana, Nevada, Utah or Wyoming. The initial data set was built using 18 expert panels that specified each wildlife species’ association with habitats and ecological conditions or variables, and that assigned confidence levels to each wildlife species for each habitat type or structural condition. The additional 12 wildlife species we added and their habitat and ecological relationships were determined during an internal review. This review re-examined the SHP database for species with similar life histories and reviewed the ICBEMP data for these 12 species. Lastly, a literature review was conducted to develop an updated life history account for each species. The literature was also used to support the depictions of how management activities were linked to the key environmental correlates data. Species nomenclature follows Collins et al. (1990), Leonard et al. (1993), and Storm and Leonard (1995) for amphibians and reptiles; American Ornithologists’ Union (1998) for birds; and Verts and Carraway (1998), Wilson and Reeder (1993), Jones et al. (1992), Hall (1981), Frost and Timm (1992) and van Zyll de Jong (1984) for mammals.
A number of primary sources were used to establish or confirm wildlife species occurrences: Csuti et al. (1997), Verts and Carraway (1998), Ingles (1965), Hall (1981), Bailey (1936), Gilligan et al. (1994), ODFW (1994), Puchy and Marshall (1993), Dvornich et al. (1997), Johnson and Cassidy (1997), Smith et al. (1997), and Brueggeman (1992). Supplemental information also came from Alexander (1996), Aubry (1982), Aubry and Houston (1992), Best (1988), and Bradley (1982).
The expert panels that developed the SHP data set had on hand range maps for each species, wildlife-habitat type and distribution maps, statewide vegetation maps, and a variety of other reference materials. These sources of material helped the panelists to determine wildlife species occurrence within a particular habitat type. To characterize the degree of association a wildlife species has with its habitat, the following categories were assigned in the SHP database:
Closely Associated - A wildlife species is widely known to depend on a habitat or structural condition for part or all of its life history requirements. Identifying this association implies that the species has an essential need for this habitat or structural condition for its maintenance and viability. Some species may be closely associated with >1 habitat or structural condition, where as others may be closely associated with only one habitat or structural condition.
Generally Associated - A wildlife species exhibits a high degree of adaptability and may be supported by a number of habitats or structural conditions. In other words, the habitats or structural conditions play a supportive role for its maintenance and viability.
Present - A wildlife species demonstrates occasional use of a habitat or structural condition. The habitat or structural condition provides marginal support to the species for its maintenance and viability.
Finally, the expert panelists assigned an overall confidence rating to the occurrence and activity headings for each species within each habitat type or structural condition. The confidence ratings were denoted as high (e.g., many peer-review published accounts), moderate, and low (e.g., few or no published accounts, mainly observations). By ascribing a confidence rating, the end user gets an idea of the overall strength of the scientific evidence.
Wildlife-habitat type is defined by O’Neil and Johnson (2001) as a group of vegetation cover types that is determined based on the similarity of wildlife use. For a detailed discussion of this approach see O’Neil et al. (1995). We used the wildlife-habitat types as defined by their approach because these habitat types can be based on current vegetation and therefore can be mapped, as well as modeled to represent historic and future conditions. Wildlife-habitat types are not species-specific because they are based on the similarity of multiple wildlife species using a suite of vegetation types, and we assume they contain the essential needs for a species' maintenance and viability. However, a wildlife species’ "habitat" refers to an individual, species-specific use of a wildlife-habitat type (Hall et al. 1997).
The wildlife habitat relationships SHP database depicts coarse-level wildlife-habitat types, structural conditions (structural and several stages of vegetation), and site-specific KECs. The Framework analysis presented in this report is based only on the coarse-level wildlife-habitat type data. As the Framework process proceeds to the subbasin levels of analysis, managers will be encouraged to integrate site-specific structure and KEC data into the subbasin-scale assessment process. The hope is that knowing the species’ relationship with its habitat type, structural conditions and KECs will help make better predictions for species occurrences and ecological conditions in an area. Knowing that ecological condition is based on physical parameters should also help to identify the key ecological functions that are operating (as well as missing) in an area. Key ecological function information for chinook and bull trout is being added to the array of wildlife in KEFs this data set as reported in the results.
We selected 3 wildlife species for conducting and exemplifying population and species-specific assessments: American black bear, bald eagle, and American beaver.
The American black bear (Ursus americanus) is widely distributed within the Columbia River Basin and is managed according to the big game or furbearer regulations in all seven states. These regulations usually allow a general hunting season and a controlled harvest to occur annually; however, most states within the Columbia River Basin have spring and fall hunting seasons. The legitimate economic value of the black bear comes from selling hunting licenses, which in turn becomes a source of revenue for individual states’ fish and wildlife agencies. An illegal economic value stems from poaching black bears and selling specific body parts (such as gall bladders) to collectors or for aphrodisiac purposes to those who highly prize their value.
Black bears occur in 32 states within the United States (Beecham and Rohlman 1994).
The black bear is a year-round resident species in the Columbia River Basin and can be found primarily associated with forested habitats that range from sea level to 8,500 feet (2590 m) (Beecham and Rohlman 1994, Vander Heyden 1997, Verts and Carraway 1998). Black bears are large mammals whose size and weight show high variability depending on food availability. Generally, adult bears range from 35 to 40 inches (89 to 102 cm) high when standing on all fours and have a length of 4 ½ to 6 feet (1.4 to 1.9 m). An adult black bear can weigh from 125 to 600 pounds (46 to 224 kg) and males are usually larger than females. The life span of black bears in the wild can be 20 to 25 years.
Most female bears breed at three years of age, but in one study in Idaho the first age of breeding was noted as 5.5 years. Females usually produce one litter with one to three cubs every two years and can be with young any time of the year. Mating occurs in June and July and the black bear has delayed implantation so the embryo does not begin to develop until November or December (Verts and Carraway 1998). Typically, the black bears den underground, in a tree cavity, or in a cave. However, they have also been known to den on the ground or in a brush pile. Females enter their dens in October or early November, and most bears leave the den in March but some females with newborn cubs may stay until April. Young are born between mid-January to mid-February and remain with the female until they are about 16-17 months old.
Sizes of home ranges vary with quality and area of habitat and with males and females. In a coniferous forest on an island in southwestern Washington, average home ranges of females were 580 acres (235 ha), and of males 1,250 acres (506 ha). In contrast, in Idaho, home ranges of males were 27,700 acres (11,210 ha) and of females 12,085 acres (4,890 ha) (Verts and Carraway 1998). These two assessments may represent the extremes in home-range size. Home range location remains relatively constant from year to year, but bears use parts of their home ranges variably depending on food availability among the seasons.
The black bear is an omnivore. Although the majority of its diet consists of grasses, forbs, berries, nuts and fruits, bears do eat mammals (e.g., elk calves), insects, carrion and fish (Jacoby et al. 1999, Berwick et al. 1986). Food is an important element of fitness as reflected in litter size, age of breeding, and in overwintering, i.e. maintaining fat reserves during hibernation (Rogers 1977). Berry crop failures have been identified as contributing to mortality of starving subadults (Jonkel and Cowan 1971, Reynolds and Beecham 1980).
Habitat studies of black bears have shown that the most important function of cover is to enable escape. Beecham and Rohlman (1994) noted that bears tended to feed near cover (<250 yd or 228 m) from a forest edge and used riparian area and stringers of timber for travel corridors as well. Sows with cubs consistently avoided clearcuts and roads, using mature timber significantly more than males. In Idaho, bears preferred to stay more than 150 ft (46 m) from roads (Beecham and Rohlman 1994). Powell et al. (1997) found that people caused a significant amount of mortality by hunting, poaching, and road kills.
The black bear is an ecological generalist whose ecological roles are important to all categories of the wildlife-habitat types where it occurs. The black bear provides 27 categories of KEFs and in some wildlife-habitat types is the only provider of some of these functions. For example, the black bear is the only identified species to physically fragment standing wood in an upland aspen forest.
The black bear is associated with 24 of the 32 wildlife-habitat types in the SHP database (Table III.B.2). As habitat generalists they do not have a close association with any one habitat type. This is further supported by its “Generally Associated” status for 18 habitat types and “Present” status for another 5 habitat types. Feeding and breeding activities in Table III.B.2 indicate the black bear feeds and breeds in 15 habitats and only feeds in eight habitat types. Confidence levels in determining these associations and activities were mostly high. However, two moderate confidence levels were noted (for coastal headlands and westside grasslands) along with three low confidence levels (for shrub-steppe and dwarf shrub-steppe).
Black bears are known and documented to have a strong and consistent association with salmon when there is an abundant population of salmon. The salmon life stages that bears associated with are spawning and when salmon become a carcass (Cederholm et al. 2000).
The HCI assessment method for the black bear is diagrammed to provide an overview of the steps taken to evaluate habitat quality across the basin (Figure III.B.2a and Figure III.B.2b). The details and code for this assessment method are outlined in Appendix F. Data were collected and analyzed for all 6-HUCs in the basin. Each 6-HUC is assigned an HCI score. Once the data were collected, we analyzed only those 6-HUCs where at least 20 percent of the 6-HUC was rated Associated (i.e., Closely or Generally Associated in Table III.B.2). If 20 percent of the HUC was rated associated, we determined if 80 percent of the HUC-6 was in the known range for the black bear. If 80 percent of the HUC was in the black bear range, we determined if 90 percent of the HUC was non-urban. If this was the case, we then determined if 50 percent or less of the HUC was in agriculture.
Sixth field HUCs that met the above conditions were assessed for three components: cover, food, and human disturbance. Cover was assessed by two variables: (1) weighted percent of the “occurrence index (Present = 1, Generally Associated = 2, Closely Associated = 3)” and (2) percent forested habitat present in the 6 - HUC being analyzed. Food was assessed by three variables: (1) weighted percent of wildlife habitat types designated as “Feeding habitat” in Table III.B.2, (2) the berry index, and (3) presence or absence of salmon carcasses in a 6-HUC. The berry index was developed for each wildlife-habitat type designated as “Feeds” or “Feeds and Breeds” in Table III.B.2. Wildlife-habitat types were given a rank depending on the number of berry-producing plant species listed in the habitat type descriptions in the SHP database (O’Neil et al. 2001). Each wildlife-habitat type habitat was assigned a high, medium, or low rank (1.0 = high, 0.5 = medium, 0.0 = low). Human disturbance was assessed with three variables: (1) Percent urban coverage in the HUC, (2) percent agricultural coverage, and (3) Road density. Results from these analyses were aggregated to determine the Habitat Condition Index (HCI) for each 6-HUC. HCI values where then aggregated up to ecological province and the entire basin. The detailed steps taken in the GIS portion of the HCI analysis are presented in Appendix F.
Throughout the Columbia River Basin, the bald eagle (Haliaeetus leucocephalus) is federally listed as a threatened species under the Endangered Species Act in all seven states. Nonetheless, because of recovery efforts lead by the U.S. Fish and Wildlife Service in partnership with other federal, state, tribes, and local governments, conservation organizations, and private entities, the bald eagle is being considered for removal from this federal designation. The delisting, a U.S. Fish and Wildlife proposal in 1999, is related to the increase in eagle numbers throughout their range. For example, in 1960 only 417 nesting pairs were found in lower 48 states and today the estimate is over 5,700. Hence, many resource biologists and managers are assessing whether or not the bald eagle warrants special protection afforded by the Endangered Species Act. This eagle is also protected by the federal law, Migratory Bird Treaty Act, as well as, under individual state’s non-game laws. The bald eagle has no legal economic value.
The bald eagle is a year-round resident in the Columbia River Basin and can be found along most major river courses and ranges from sea level to 8,000 feet (2438 m) (Garrett et al. 1993; Stalmaster 1987). Adults weigh seven to 10 pounds (2.6 to 3.7 kg) with a wingspan of 6 ½ feet (2 m). They have been known to live more than 20 years in the wild, and the age at first breeding is usually five years. The bald eagle builds its nest in the tops of large live trees usually near water and nests can be up to 20 feet wide (6 m) and weigh up to 4,000 pounds (1492 kg) (Knight et al. 1983, Hermata 1989). Food habits vary with seasons and locations, and they do take advantage of fish (suckers, trout, and whitefish), birds (particularly waterfowl), salmon carcasses (especially in the fall and winter), and mammal carrion (Stalmaster 1984 and 1994).
Bald eagles begin laying eggs from early March to early April with mean hatching dates from mid-April to mid-May. Incubation of eggs usually lasts 34-36 days and they fledge one to two young per year but can fledge three on occasion (Stalmaster 1987). Fledging occurs 10 to 14 weeks after hatching or generally in early August. Bald eagles usually have some fidelity with the nest sites and even though they can travel great distances (Gerrard et al. 1978), most nest within 100 miles of where they were originally raised (Jenkins and Jackman 1993). Bald eagles are known to use a communal roost especially in the winter when salmon are spawning. Up to 300 bald eagles may use a single roost site (Knight et al. 1983).
Human disturbance can affect perching, roosting, and feeding (Fraser 1985, Knight and Knight 1986). Bald eagles are more sensitive to human activities on the river (boating or fishing) than to vehicle traffic or airplane flight (Stalmaster and Newman 1978, Knight and Knight 1984, and Department of Interior 1986). However, bald eagles can tolerate some human activity where there is an abundant food supply and adequate habitat (Stalmaster and Newman 1978; Steenhof 1978). Human activity that occurs beyond 1/3 of a mile (or 500m) from a bald eagle use area seldom disturbs the birds (Stalmaster and Newman 1978).
The bald eagle is associated with 23 of the 32 wildlife-habitat types (Table III.B.1) listed for the region. Associations with wildlife habitat types in Table III.B.1 indicate the bald eagle is “Closely Associated” with one habitat type, open water; “Generally Associated” with 16; and “Present” in another eight. Feeding and breeding activities in Table III.B.2 indicate the eagle just breeds in nine, breeds and feeds in an additional five, and just feeds in another 11. Confidence levels with making these determinations are mostly high with the exception of one wildlife-habitat type, feeding in westside grasslands, which is low.
The bald eagle has a strong and consistent relationship with salmon as a predator on salmon. This relationship extends to three salmon life stages: saltwater residence (when they are smolts, immature, and adults), spawning, and carcasses. The eagle also has an indirect relationship with salmon because they are known to feed on birds that also feed on salmon (Cederholm et al. 2000).
The bald eagle model, similar to the black bear model, was developed to assess habitat quality (i.e., Habitat Condition Index) for three time periods: historic, current and future. Input variables selected for the model were based on: (1) availability of a consistent data set for all 6-HUCs in the basin, (2) importance to bald eagle nesting, roosting, and foraging and (3) likelihood that proposed management activities would influence the variables. Our review of the literature (summarized above) indicated that consistent information on nesting and roosting sites was not available for all 6-HUCs in the basin. Consequently, nesting and roosting sites and the influence of human disturbance on them was not included in the model. Analyses at the subbasin level are more likely to have access to local databases and professionals who are familiar with nest and roost sites and human use areas.
The Habitat Condition Index for the bald eagle was developed by evaluating generalized foraging and breeding information for the various wildlife-habitat types in the SHP database. Habitat associations (Closely and Generally Associated in Table III.B.1) and habitat activities (feeds, breeds and feeds, and breeds in Table III.B.1) for the various wildlife habitat types in the SHP database were the main input variables we used to evaluate food and cover, as outlined in Figure III.B.3. The detailed steps taken in the GIS portion of the HCI analysis are presented in Appendix F.
The American beaver was selected as a species to assess in the Multi-Species Framework because of the association with aquatic ecosystems and with fish diversity and abundance (Schlosser and Kallemeyn 2000). Our HCI assessment of the American beaver included three components: physical condition, cover, and food, similar to our assessment of black bear and the bald eagle. The first step of the of analysis steps, data collection, was problematic for this species. The scale of the habitat data was too coarse and the assessment method met with failure as discussed in the Results.
The American beaver (Castor canadensis) is widely distributed within the Columbia River Basin and is govern by the furbearer regulations that are in place in all seven states. The legitimate economic value of the beaver comes to individuals who trap them and in turn sell their pelts. Sometimes, the beaver is considered a nuisance because it can cause erosion, blockages or flooding. Hence, landowners can have beaver removed or relocated depending on the amount of damage being sustained. The American beaver occurs in all 50 states (Hill 1982), and the estimated population in the early development of North America is 60,000,000.
The beaver is associated primarily with forested and aquatic habitats from sea level to 7,500 feet (2286 m) elevation (Verts and Carraway 1998). Beavers’ size and weight show high variability depending on food availability. Adult beavers are up to 47 inches (120 cm) long and weigh from 47 to 83 pounds (16 to 31 kg). The lifespan of beavers in the wild can be greater than 20 years, however few live beyond 10 years (Jenkins 1979).
Most female beavers breed at three years of age, the average number of offspring per litter is three, and they only have one litter per year. Mating occurs mostly in January through February and the gestation period takes 105 - 107 days. Most of the offspring are born in May and June, and the young are weaned at two to three months and leave the natal lodge at the end of their first or second year (Verts and Carraway 1998). Typically, beavers build lodges on banks of streams or ponds, or burrow in banks.
Beavers usually form colonies consisting of a mated pair, their yearlings, and offspring of the year. They have variable dispersal distances from 1.8 – 13.8 miles (2.9 - 22.2 km) and dispersers of various ages averaged 5.6 miles (9.0 km) in a straight line distance (Leege 1968). Beavers are active all year long.
A beaver colony is a single group of four to eight animals per stream reach. A colony uses a common food supply, and maintains common dams. An average of one to two colonies/mile of stream occur in good habitat (Lawrence 1954; Aleksiuk 1968). Naiman et al. (1986) suggested that beavers are “keystone” species because of their relationship to salmonids and ability to “affect ecosystem structure and dynamics far beyond their immediate requirement for food and space.” Removal of beavers has been shown to fundamentally alter aquatic ecosystem functions (Spence et al. 1996).
The American beaver is associated with 16 of the 32 wildlife-habitat types in the SHP database (Table III.B.3). As an aquatic specialist it is “Closely Associated” with four open water, wetland and riparian habitat types (Table III.B.3). It is “Generally Associated” with eight habitat types and “Present” in another five habitat types. The beaver is identified to feed and breed in four wetland habitats and in subalpine parkland where wet meadows occur (Table III.B.3). Confidence in these associations and activities is mostly high. However, moderate confidence levels were noted for agriculture, urban, and subapline parklands, along with one low level of confidence for western juniper and mountain mahogany woodlands).
The beaver is not known to eat or prey on salmon. However, from a habitat standpoint, the beaver does have a close association with salmon because of its ability to create ponds and enhance functional processes that are favorable for salmon (Hill 1982; Schlosser and Kallemeyn 2000).
Input variables to the beaver HCI model included stream hydrology, wildlife-habitat types, food, and breeding. Habitat types, food, and breeding were assessed using information from the SHP database (Table III.B.3). Also, the value of the agriculture habitat type was discounted in the model even though beaver can occur in this habitat type. The stream hydrology conditions that were assessed included monthly flow and amount of sinuosity (meandering in a stream). A diagram of the model outlining the steps of the analysis process is not presented for the beaver because the site-specific stream hydrology data and fine-scale wildlife habitat data were not available on a consistent basis across the basin. Details of the proposed analytical method to assess beaver HCI are summarized in Appendix F.
Two wildlife-habitat type maps were developed for the Multi-Species Framework process to depict historic (potential) and current conditions. These maps served as a base for making assessments between the historic and current conditions. Future (year 2100) conditions under each alternative were also developed (Vail, Appendix L). Collectively, we used the two maps and the future habitat types projection to evaluate black bear, bald eagle, beaver, and key ecological functions of wildlife species under historic, current, and future conditions.
Two wildlife-habitat type maps were developed for the Multi-Species Framework process to depict historic and current conditions. These maps served as a base for making assessments between the historic, current, and alternative conditions for evaluating black bear, bald eagle, beaver, and key ecological functions. Wildlife-habitat type maps are useful for integrating concepts so that the outputs can be visually displayed. However, there are some limitations in their use as discussed below. Comparing two maps can show how vegetation communities can change through time. The Interior Columbia Basin Ecosystem Management Project and related assessments also address vegetation changes since early historic times, as reported by Everett et al. (1994), Hann et al. (1998), Hessburg et al. (2000), and Huff et al. (1995).
The Northwest Habitat Institute (NHI) developed a map depicting the current distribution of the 32 wildlife habitats types, described by the SHP project, for the Columbia River Basin in the United States. This map was compiled from existing vegetation maps that were created for each state as part of the National Gap Analysis Program sponsored by US Geological Survey, Biological Resource Division (USGS/BRD). Each state’s map is based on interpreting vegetation cover data from satellite imagery. Vegetation maps from all or parts of seven states (Idaho, Montana, Nevada, Oregon, Utah, Washington, and Wyoming) in the Columbia River Basin were used by NHI to develop the wildlife habitat types map depicting current conditions.
The primary purpose for developing the vegetation maps for the National Gap Analysis Program was for USGS/BRD to conduct statewide biodiversity assessments. Hence, the resolution of their vegetation maps reflects a statewide, regional, or coarse resolution for planning. That is, their maps can serve as an initial basis for large-scale mapping or database investigations but they are more accurately interpreted at the statewide or province scales, and only for some of the largest subbasins.
Hence, the current wildlife-habitat type map provides only an initial depiction of the amounts of wildlife habitats that may exist within watersheds, but is not of sufficient resolution for depicting the site-specific location of habitats within each watershed. The minimum mapping unit for the basin-wide map is 250 acres (100 ha), whereas a more appropriate scale for within watershed assessments would be 10-75 acres (4-30 ha) depending on land ownership and habitat patch sizes. Thus, wildlife habitats that occur in patch sizes less than 250 acres, e.g. linear riparian habitat, are likely underrepresented in the current map.
Further, there has been no formal
validation of the basin-wide current wildlife habitat map.
Because maps are only a representation of reality and cannot depict all
the detail represented in nature, some generalization is unavoidable.
Remotely sensed maps developed from photo interpretation or satellite
imagery also contain some errors. Conducting
an accuracy assessment allows the user to know at a glance what the overall
reliability is, so that when decisions are made the accuracy of the map can be
taken into account. Because of
the size of the mapping area, time frame, and costs, no formal accuracy
assessment was done. However, the
National Biodiversity Gap Analysis Program had a goal of 80 percent overall
accuracy for each state’s vegetation map, and NHI accepted their stated
validity of their map products.
Finally, because there is a desire
to move towards subbasin information, which would entail maps produced at
finer resolutions than presented in this report, accuracy assessments may be
less critical or a lower priority for the current array of map products than
for later map products produced at the subbasin scale.
We do recognize the importance of conducting accuracy assessments and
that they would be critical to the utility and acceptance of subbasin-scale
maps as a tool for resource managers. In
general, accuracy assessments would entail determining the classification
error in maps by using an a priori
target level of thematic map accuracy (for subbasin mapping we would propose a
per class accuracy of 75 percent and overall map accuracy of 80 percent) and
designing the empirical assessment (number of sampling points, etc.) based on
statistical sampling procedures (Stehman, 1992).
A historic (potential) map was developed by NHI by combining products from two previous works: Interior Columbia Basin Ecosystem Management Project (ICBEMP; USDA Forest Service 1997), and the Oregon Biodiversity Project (Defenders of Wildlife 1998). These two mapping efforts used very different methods. The ICBEMP historic data were mostly derived from a model, whereas at least a portion of the Oregon Biodiversity Project map was created by using surveyors’ notes from the 1850 land survey.
NHI combined these efforts to create a wildlife habitat map that depicts historic (potential) conditions of the Columbia River Basin in the U.S. The result is a historic map that is a theoretical construct with a coarse (1-km square pixel size) level of resolution designed to give a regional perspective. This map can provide only initial approximations of the presence and distribution of wildlife habitat types within specific subbasins and watersheds because of the need for more detailed information at these levels.
Because of the limitations with the historic map, no validation of this map was done. We are unaware of any previously collected detailed information for all the subbasins and watersheds throughout the specific geographic areas of basin addressed in this project. Further, because there are no recognized historical data sets that would give such a basin perspective, validation would be difficult. Hence, the historic map best depicts gross generalizations of gains or loses of specific wildlife habitat types. Additionally, it can give a user an idea of what the potential may have existed within provinces and within larger subbasins.
Many species of wildlife are affected by both the general macrohabitat conditions, depicted in our maps as wildlife habitat types, and by the specific structure of vegetation. However, to accurately depict distribution and abundance of vegetation structure would require spatially explicit data sets at both coarse and fine levels of resolution. A coarse level map typically has a minimum mapping unit (mmu) of about 250 acres, whereas a finer level map shows details at about a 10-acre mmu.
Because fine-level data are either not available or have not been synthesized for all lands basin-wide, the outcomes presented here should be used to interpret wildlife-habitat type information only at a coarse scale. Vegetation structural conditions are best depicted at a finer level of resolution, that is, at a stand level with a 10-to-40 acre mmu, and should be included in future subbasin mapping efforts.
Consideration of vegetation structure can greatly influence analysis and interpretation of wildlife-habitat relationships. We selected a few subwatersheds (6th HUC) for which vegetation structure information was available, and found that consideration of structural condition influenced results of projecting wildlife species and their key ecological functions within the area. Thus, we concluded that, at such finer scales of mapping resolution, vegetation structure likely influences the presence and distribution of wildlife species and thus overall ecosystem biodiversity, productivity, and sustainability.
The ecological approach we adopted for the Multi-Species Framework supplements the emphasis on coldwater fish with tools that address ecological functions of all wildlife and eventually all fish in the Columbia Basin. The ability to address and describe in a repeatable way ecological functions of all vertebrates (including humans), using a common database, is a new approach to broad-scale resource assessment presented by the Multi-Species Framework.
The term key ecological functions (KEFs) of wildlife refers to the principal set of ecological roles performed by each species in its ecosystem (Marcot and Vander Heyden, 2001). KEFs refer to the main ways organisms use, influence, and alter their biotic and abiotic environments. “Key” refers to the main roles played by each species. Categories of KEFs can be depicted for each species and used in multiple-species analyses of alternatives for land management in the Columbia River Basin.
One major assumption on this analysis is that wildlife KEFs contribute to and affect ecosystem biodiversity, productivity, and resource-use sustainability (BPS). Another assumption is that the parameters of BPS describe ecosystem integrity, the maintenance or restoration of which, we presume, can be one prime goal of ecosystem management. The purpose of tracking wildlife KEFs, including their patterns and changes, therefore, is to determine how management actions might affect wildlife, the biotic functioning of ecosystems, and ecosystem BPS. It serves as a way to measure the degree to which ecosystem management goals are met for maintaining or restoring at least some facets of ecosystem integrity.
We measure changes in wildlife KEFs in several ways, including historic and current patterns of, and future potential changes in: (1) the distribution and abundance of species, based on their habitat associations, that perform particular ecological roles (that is, that are coded for particular KEF categories); (2) Functional redundancy of KEF categories; and (3) The richness and diversity of KEF categories that ecological communities can support. Functional redundancy refers to the number of species performing a particular KEF category. As stated by Marcot and Vander Heyden (2001), the basic premise is that functionally redundant, rich, and diverse communities may be more resistant or resilient to adverse disturbances (MacNally 1995, Naeem 1998, Rastetter et al. 1999) and can more consistently support greater levels of biodiversity (Jaksic et al., 1996 Walker 1992) than can less functionally redundant, rich, or diverse communities.
Marcot and Vander Heyden (2001) noted that ecological implications of functional patterns of species and communities, and their influence on BPS, can be taken as testable hypotheses about the roles of wildlife and how ecosystems work. They listed several key such hypotheses, with perhaps the most important ones for the current work being:
1.
Functional redundancy imparts community resilience: for a particular
function, the higher the functional redundancy, the greater ability of the
community to resist stresses put on that function.
2.
The greater the functional redundancy, the more sustainable is the set
of resources that the function provides.
3. The more functionally rich and diverse a community, the greater is its natural productivity and its native biodiversity.
Over time, such hypotheses could be tested in the context of adaptive management by comparing performance of BPS over time or among areas managed differently.
Collectively, the methods we used provide a means of determining the degree to which an ecosystem is “fully functional,” by comparing historic, current, and potential future KEF conditions. Fully functional ecosystems are those that have the full set of historic KEF categories, and the historic patterns of functional redundancy for each KEF category.
Other functional aspects include determining: (1) Functional
richness, which is the number of KEF categories performed by species in a
community, (2) Total functional
diversity, which is functional richness weighted by functional redundancy
(Brown 1995), analogous to species diversity, and (3) Functional
web, which is the full array of all KEFs associated with a set of species
that may be specified by some habitat element or structure (Marcot and Vander
Heyden 2001). Because many
functions can extend beyond a habitat element or structure, functions that are
supported in part by specific KECs can influence parts of the ecosystem well
beyond those KECs. For example,
the ecological functions provided by beaver extend well beyond the confines of
the KEC of water depth (Schlosser and Kallemeyn 2000).
Scientific and common names and species occurrence status, by state, were reviewed by Dick Johnson (Washington State University), B. J. Verts (Oregon State University), Tom O’Neil (Northwest Habitat Institute), Rolf Johnson (Washington Department of Fish and Wildlife [WDFW]), Derek Stinson (WDFW), Kelly Bettinger (WDFW), Charlie Bruce (Oregon Department of Fish and Wildlife [ODFW]), Kelly McAllister (WDFW), Bruce Mate (Oregon State University Marine Science Lab), Steven Jeffries (WDFW), and Robin Brown (ODFW). Taxonomic order follows regional publications or commonly accepted national books to facilitate cross-referencing.
The Species Habitat Project (SHP) assigned five occurrence status categories to each wildlife species in the SHP database: occurs, accidental, introduced, reintroduced, and extirpated; the species could be listed as any one of these categories in any state within the basin (Johnson and O’Neil 2001). Occurs means >15 documented observations, that is, they are considered to be common species for the area. Some species listed as “occurs” do not have 15 records in recent decades, so there are species listed that were formerly more abundant, but now may be considered rare (like the short-tailed albatross). This figure of 15 documented observations was derived from its use in the states’ ornithological groups, such as the Oregon Field Ornithologists. Accidental denotes those species with <15 documented occurrences, or >15 records but the Columbia Basin is not a regular part of the species’ range. Introduced denotes species that are not native (that is, that likely did not occur before European settlement) but that now breed in the Columbia Basin. Reintroduced denotes native species that were eliminated from the Columbia Basin or reduced to such low population levels that additional individuals were required to supplement or re-establish the species. Extirpated refers to a native species whose originally native populations have been completely extirpated from the Columbia Basin.
Three categories were used to describe the breeding status of the species. Breeds is for those species with >5 documented breeding records by separate pairs unless professionals familiar with the species believed that breeding is probable but has not yet been documented. Non-breeder refers to those species that occur in the state(s) but do not breed, or have <5 documented breeding records. Bred-Historically refers to those species that used to breed in the state(s) but currently do not.
NHI did the alternative strategies analysis for the black bear and bald eagle using basin-wide species distribution maps. NHI used data from previous inventories or studies to validate these basin range maps. For example, 29 years of bald eagle inventory data helped determine which 6th HUCs should be a part of the bald eagle’s basin-wide range. With black bear, the radio locations from a 3-year study in the Central Cascades of Oregon helped clarify the Habitat Condition Index by comparing NHI’s black bear distribution map with that from the study. Additionally, species range information from Idaho, Oregon, Western Montana, and Washington further helped with the validation of NHI’s distribution maps of bald eagle and black bear by comparing the maps with the Habitat Condition Index model’s ranges. Our last review for accuracy compares differences between species ranges in the literature with known occurrence of suitable habitat. The databases and analyses of wildlife KEFs were not validated; nor were the patterns of community function resulting from our Ecological Functions Analysis. Clearly, work remains to better quantify the ecological roles of wildlife, how those roles affect BPS, and how management actions affect the functioning of communities.