Habitat Evaluation Procedures (HEP) Report; Sandy River Delta, Technical Report 2000-2002.

Land managers are often challenged with the mandate to control exotic and invasive plant species. Reed canarygrass (Phalaris arundinacea) and Himalayan blackberry (Rubus discolor) are 2 such species that are currently threatening natural areas in western United States. Reed canarygrass may be native to the inland northwest (Antieau 2000), but it has invaded many wetland areas as dense, monoculture stands. Spread of this plant species is largely attributed to human disturbances, e.g., draining, farming (Antieau 2000). Reed canarygrass often dominates other emergent vegetation such as cattail (Typha spp.) and bulrush (Scirpus spp.) (Whitson et al. 1996, Apfelbaum and Sams 1987), and the resulting habitat is largely unsuitable for wetland birds. Himalayan blackberry was introduced to the United States as a garden shrub and was planted at wildlife-management areas for food and cover. It easily colonizes disturbed places, such as roadsides, ditches, and flood plains (Hoshovsky 2000). Once established, it forms a thick, impenetrable stand, which excludes native shrub species. Although Himalayan blackberry does provide food and cover for wildlife, particularly during fall and winter, it decreases habitat diversity, and therefore, may decrease wildlife diversity. Furthermore, patterns of avian nest predation may be altered in some exotic-shrub communities (Schmidt and Whelan 1999). For land managers to make sound decisions regarding invasive-plant control, it is useful to obtain information on current plant distributions in relation to targeted wildlife species, and then use models to predict how those species may respond to changes in vegetation. The Habitat Evaluations Program was developed by the U.S. Fish and Wildlife Service to evaluate current and future habitat conditions for fish and wildlife (Stiehl 1994). The program is based on Habitat Suitability Index (HSI) models for specific wildlife species. Each model contains several variables that represent life requisites (e.g., food and nesting cover) for that species. These variables are evaluated with vegetation sampling, and/or through the interpretation of aerial photographs and the like. Variable values are assigned a numerical score. The score may be based on a categorical rating (e.g . , different vegetation types receive different scores based on their importance for that species) or may be the result of a linear relationship (e.g., the score increases with the variable value; Figure 1). Variable scores are then input into a mathematical formula, which results in an HSI score. The HSI score ranges from 0-1, with 0 representing poor-quality habitat and 1 optimal habitat. HSI models assume a positive, linear relationship between wildlife-species density and the HSI score. For example, with an HSI score of 1, we assume that a species will be present at its highest density. Models can be projected into the future by changing variable values and observing the corresponding changes in HSI scores. Most models are relatively simple, but some are complex. These models have come under considerable scrutiny in the last several years, particularly concerning the validity of model assumptions (Van Horne 1983, Laymon and Barrett 1986, Hobbs and Hanley 1990, Kellner et al. 1992). Regardless of criticisms, these models may be used with success when there is an understanding and acceptance of model limitations. Each model should be evaluated as to its applicability in a given situation. Model validation, where results have on-the-ground verification, is highly recommended. Specific objectives of this project were to (1) conduct avian surveys and measure the present vegetation at the Sandy River Delta, (2) input the vegetation data into HSI models for 5 avian species, (3) evaluate the current habitat suitability for these species, and (4) predict species responses to potential changes in vegetation, resulting from the removal of reed canarygrass and/or Himalayan blackberry