Considering Habitat Suitability Index Models for Avian Habitat Design

Summer 2013 Ecesis, Volume 23, Issue 2

The Nature Conservancy’s Cosumnes River Preserve located in southern Sacramento County consists of approximately 46,000 acres of natural and agricultural lands adjacent to the Cosumnes and Mokelumne Rivers. Habitats include riverine, riparian and oak woodland, willow and buttonbush scrub, natural and managed wetlands, annual grasslands, vernal pools, and agricultural lands. The preserve provides monthly opportunities for the public to participate in the monthly Tall Forest bird survey. This event is the only public access to a location TNC describes as “a woodland that bears great resemblance to pre-Anglo Central Valley riparian forests, a habitat that is all but gone.” (

I participated in a Tall Forest survey several years ago and, drawn by the allure of what I had observed before and the opportunity to go birding in the Tall Forest, I participated in the June 2013 survey. This article was inspired by my visit to the Tall Forest and my interest in wildlife, particularly birds. I’ve spent hours flipping through field guides to memorize field marks and habitat requirements and many, many more hours scanning, searching, and skulking for birds everywhere I go.

As restoration ecologists, our goal is often to restore habitat to meet mitigation acreage requirements. As a life-long nature enthusiast, biologist, and avid birdwatcher, I enjoy observing and learning about different vegetation communities, which plants and wildlife species occur there, and what the habitat quality of a given community is for those species. When I visit a restoration project site, particularly during the post-project performance monitoring and establishment period, I look through this lens and ponder not just whether or not we’ve met the performance standards and permit requirements, but if this fledgling habitat is providing or is trending toward providing the intended habitat quality goals for resident and migratory species. Sometimes I could swear I hear the birds say, “Is this the best you can do?”
I think we have all reflected, and know, that in most cases we cannot truly restore native ecosystems because we cannot restore pre-disturbance landscape processes. Askins (2002) defines habitats that have been almost completely destroyed or that have been dramatically transformed as lost landscapes. Unfortunately we often have to rely on these lost landscapes as our reference sites and in most cases it is beyond our control to restore landscape processes. So we must carry on as President Theodore Roosevelt stated, “Do what you can, with what you have, where you are.”

During his presentation at the 2013 SERCAL conference (Assessing Functional Lift for Floodplain Restoration Alternatives Using Ecohydraulic Modeling), Andy Collison of ESA PWA pointed out that habitat mitigation has historically placed more emphasis on the acreage of habitat restored than the function of that habitat, but that this paradigm is shifting with the emphasis on functional lift.

Habitat suitability index (HSI) models offer a tool to guide restoration design for optimal habitat conditions for target species and other species in the same guild — to the extent practicable. HSI models are used in Habitat Evaluation Procedures (HEP) to estimate the value of the habitat in a project area for the selected evaluation species (Stiehl 1995). 

HSI models have been developed for everything from the American oyster to moose. Models are constructed from basic life history information for an evaluation species or by modifying existing habitat models based on regional or local habitat conditions and life history requirements (USFWS 1981). Models can be developed for a specific evaluation species (e.g., yellow warbler) or for land cover types. A HEP assessment may employ one or more HSI species models to capture the different habitats present or proposed. This article will focus on species models.

Species models define the geographic area to which the model is applicable (e.g., entire breeding range of the species, regional ranges), describe the habitat requirements needed by the species (e.g., food, nesting, cover), identify the key habitat variables for a particular life requisite (e.g., density and height of cover required for nesting), and identify a range of suitability index scores from 0 to 1.0, with a value of 1.0 representing the optimal suitability of a given variable. 

As with any models, there is debate about the accuracy and reliability because thorough validation and verification studies have not been performed to evaluate model efficacy (Tirpak et al 2009, Roloff and Kernohan 1999). Some models are considered invalid because of sampling in a limited range of habitat conditions and inadequate population sampling. This may be the case for models that cover a broad geographic area, such as a yellow warbler model that applies to the entire breeding range of the yellow warbler (Schroeder 1982). However those models that are developed for regional populations may be more accurate because of the smaller sampling area and presumably less variation in the range of habitat conditions. 

The U.S. Fish and Wildlife Service’s Sacramento Field Office developed an HSI model for rufous-sided towhee (USFWS 1984). In California, the Rufous-sided towhee — since split into two species, the Spotted towhee of the western U.S. and the Eastern towhee — inhabits areas with dense shrub cover, including riparian and oak woodland, chaparral, and thickets, among others. Although the HSI model for towhee cited herein is a draft (a final was not prepared), I selected Spotted towhee for this example. It is one of the most common birds in the Tall Forest because the habitat there is optimal for towhee. The HSI habitat variables and the optimal habitat conditions (suitability index score of 1.0) for towhee are shown in Table 1. 

Spotted towhees are abundant in the Tall Forest because the dense understory provided by wild grape, California wild blackberry, poison oak, shrub species, and small trees provide optimal cover, nesting, and foraging habitat. The Tall Forest is subject to seasonal inundation that supports these conditions. Depending on location, habitat that supports resident towhees also provides breeding and foraging habitat for numerous other resident species including House wren, Bewick’s wren, and California towhee, and wintering species including Fox sparrow and Hermit thrush. Although optimal habitat for towhees may not represent optimal habitat for these other species, considering the model variables and developing restoration plans and assessing landscape processes that will provide optimal habitat for towhees will benefit at least other ground dwelling species that rely on dense cover — both birds and other wildlife. 


As shown in Table 1 (right), towhees prefer dense cover and require a leaf litter and humus layer. When I look at many young riparian and oak restoration sites and observe a lack of floodplain connectivity and the relatively dry soil conditions that support grasses instead of dense thickets of vines and shrubs, I wonder whether or not these areas will be capable of developing into high-quality habitat for towhees and other wildlife species dependent on this dense understory habitat. It may be possible to establish dense understory habitat by using a high-density planting shrub, vine and forb species and more extensive maintenance and management actions. However, so many mitigation projects are subject to just meeting the permit requirements for acreage. There is also typically a limited maintenance and monitoring time-frame and funding is usually a factor.

Unless developed for a very specific location, the habitat variables and suitability of any HSI model cannot be completely representative of local conditions. However, I think there is a benefit in using HSI models, in combination with field assessment of reference and project sites. Knowing what bird species occur in the vicinity of a proposed restoration site and understanding the critical habitat variable and associated suitability indices can be useful design tools for habitat design, restoring or improving landscape processes, and developing appropriate plant palettes and seed mixes to benefit target species.

There are numerous sources of HSI models, including the USFWS ( and U.S. Geological Service ( The California Department of Fish and Wildlife also maintains a document database of HSI models for use with ARC/INFO. There are also numerous regionally developed models online. 

Nothing can replace good old-fashioned field work, but the information available in HSI models can be a useful information source. — by Harry Oakes, Senior Manager, ICF International, and SERCAL Region 2 Director

Literature Cited:
Askins, Robert A. 2002. Restoring North America’s Birds. Lessons from Landscape Ecology. Second Edition. Yale University Press.
Garrison, B.A. Draft Habitat Suitability Index Model: Northern Oriole (Icterus galbula). September. U.S. Fish and Wildlife Service, Division of Ecological Resources. Sacramento, California.
Roberts, R. Chad. 1986. Habitat Suitability Index Model. Riparian Songbird Guild. Humboldt Bay, California. August. Prepared for California State Coastal Conservancy. Prepared by Oscar Larson & Associates. Eureka, California.
Roloff, G. J.and B. J. Kernohan. 1999. Evaluating Reliability of Habitat Suitability Index Models. Wildlife Society Bulletin 27: 973 – 985.
Stiehl, Robert B. 1995. Habitat Evaluation Procedures Workbook. January. Prepared by the National Biological Service Midcontinent Ecological Science Center. Fort Collins, Colorado.
Schroeder, Robert L. 1982. Habitat Suitability Index Model: Yellow Warbler. July. Prepared by the U.S. Fish and Wildlife Service, Fort Collins, Colorado.
Tirpak, John M. et al. 2009. Assessing Ecoregional-Scale Habitat Suitability Index Models for Priority Landbirds. Journal of Wildlife Management, 73(8):1307 – 1315.
U.S. Fish and Wildlife Service. 1981. Ecological Services Manual – Standards for the Development of Habitat Suitability Index Models. 103 ESM. April. Prepared by the U.S. Fish and Wildlife Service. Washington, D.C.
U.S. Fish and Wildlife Service. 1984. Draft Habitat Suitability Index Model Rufous-Sided Towhee (Pipilo erythrophthalmus). September. Prepared by the U.S. Fish and Wildlife Service. Sacramento, California.