the role rbance in thecology and ofbirds* · silviculture stands apart as a commercially driven...

._. 1._

AJo_4,,,,.;1,_kle

Annu. Rev. Ecol, SysL 2001.32:251-76

THE ROLE OF DISTURBANCE IN THE ECOLOGY

ANDCONSERVATION OFBIRDS* z "

..g_Jeffrey D. Brawn, 1Scott K. Robinson, 2 _"and Frank R. Thompson III3 _"_i _ _tIllinois Natural History Survey and Department of Natural Resources and EnvironmentalSciences, Universityof lllinois, Champaign,Illinois 61820; e-mail.'[email protected] _ _ _ _o:Department ofAnimal Biology, University of Illinois, Champaign, Illinois 61820; _ _ __ _"_"e-mail."[email protected] _" _3 _ o3USDAForestService, North CentralResearch Station, Universityof Missouri, Columbia, _ _: _ _Missouri 65211-7260; e-mail:[email protected] _ _ _" _

O_ _ 0 _'.Key Words flood pulse, fire, habitat selection, silviculture, successional habitats o_'_ o-- _

_0_• Abstract Natural ecological disturbance creates habitats that are used by diverse ._ = _"<; 0_ _-_groups of birds. In North America, these habitats or ecosystems include grasslands or _o_--_ _prairies, shrublands, savannas, early successional forests, and floodplains. Whereas the _ _ o _extent of all natural habitats has diminished significantly owing to outright loss from _ =" = _

agriculture and development, the suppression of disturbance by agents such as fire and _ _ i 7_flooding has led to further losses. Accordingly, the abundances of many bird species _ __'_adapted to disturbance-mediated habitats have declined as well. In North America, - _ "

these declines have been more severe and common than those of species associated -q . _.GX

with less frequently disturbed habitats such as mature or closed-canopy forests. Fieldstudies consistently reveal the direct role of disturbance and successional processesin structuring avian habitats and communities. Conservation strategies involving themanagement of disturbance through some combination of flooding, application of fire,or the expression of wildfire, and use of certain types of silviculture have the potentialto diversify avian habitats at the local, landscape, and regional scale. Many aspects ofthe disturbance-ecology of birds require further research. Important questions involveassociations between the intensity and frequency of disturbance and the viability of birdpopulations, the scale of disturbance with respect to the spatial structure of populations,and the role of natural vs. anthropogenic disturbance. The effects of disturbance andensuing successional processes on birds are potentially long-term, and comprehensivemonitoring is essential.

*The US Government has the right to retain a nonexclusive, royalty-free license in and toany copyright covering this paper.

251

252 BRAWN • ROBINSON • THOMPSON

INTRODUCTION

Disturbance was once viewed largely as an insult to the "balance of nature" and

synonymous with habitat destruction (see Botkin 1990). Certain forms of distur-bance, however, are now held by ecologists and conservation biologists to play afundamental and creative role in maintaining the natural heterogeneity in envi-ronmental conditions that organisms experience through space, time, or both.

Much theory and empirical work has been devoted to understanding how suchheterogeneity, or patchiness, affects the evolution of life histories and key ecolog-ical processes at the population, community, ecosystem, and landscape levels oforganization (Connell 1978, Pickett & White 1985, Southwood 1988, Alversonet al. 1994). The importance of disturbance to the ecology of species and conser-vation of biodiversity has gained widespread recognition (Cormell 1978, Sonsa1984, Pickett & White 1985, Petraitis et al. 1989, DeGraaf & Miller 1996a,

Askins 2000) and has been defined by Pickett & White (1985, p. 7) as "anyrelatively discrete event in time that disrupts ecosystem, community, or popu-lation structure and changes resources, substrate availability, or the physicalenvironment."

In the 1970s and 1980s, a series of models appeared that related disturbance

to enhanced species diversity at the local (i.e., within-patch) scale. The positedmechanism was generally the prevention of competitive dominance and equilibrialconditions either directly (through density-independent mortality of organisms) or

indirectly (through changes in habitat and resource levels). Empirical evaluationof these models has been generally corroborative, but the majority of these studies

have been on space-limited, sessile organisms such as barnacles, mussels, andplants (Sousa 1984).

More recent works have considered the role of natural disturbance on species

diversity at the landscape or regional scale (Angelstam 1998, Askins 2000). Now,disturbance is viewed as a natural ecological process leading to a mosaic of habitatsor successional stages that may enhance both u and/_ diversity (Anglestam 1998).Indeed, ecological disturbance is an assumption in models that consider spatially

structured populations. Depending on the scale of disturbance and the scale atwhich an organism "perceives" the environment (Vos et al. 2001), natural and,increasingly, anthropogenic disturbance can form the landscape framework for

metapopulations and source-sink dynamics (Hanski 1991, Dunning et al. 1992).At a more evolutionary time-scale, the frequency of habitat disturbance has been

hypothesized as the source of selective pressure underlying the evolution of dis-persal strategies and other key life-history Waits (Southwood 1988). In restorationecology, landscape ecology, and the concept of ecosystem management, naturaldisturbance is now generally recognized as essential for maintaining biodiver-

sity (Alverson et al. 1994, Askins 2000). Direct application of disturbance (e.g.,prescribed burning or silviculture) or allowing the natural agents of disturbance(e.g., wildfire or floods) to proceed, are becoming standard elements of local andregional conservation strategies (Johnson et al. 1998).

DISTURBANCE ECOLOGY AND BIRDS 253

This paper reviews how natural and anthropogenic disturbance can affect thepopulation and community ecology of birds, and it emphasizes the conservationimplications of these effects. The motivation for this review is twofold. First,the ecology of disturbance with relatively mobile organisms is different, but lesswell understood, than that for sessile organisms (Sousa 1984). Second, evidence

is growing that some form of disturbance is required by the habitat needs of alarge segment of the world's terrestrial avifauna (Askins 2000). Much has beenwritten on the loss and fragmentation of closed-canopy forests due to agricultureor urbanization (e.g., Robinson et al. 1995), but less study has been devoted to therole of managing ecosystems and successional processes through the application ofdisturbance. We limit our review to bird species associated with terrestrial habitatsor ecosystems and emphasize species that breed in North America. Finally, wefeature the major agents of disturbance that we believe are most relevant to land-usepolicy or options for avian conservation: fire, silviculture, and floodplain dynamics.Silviculture stands apart as a commercially driven source of habitat disturbance;however, its effects are pervasive and potentially creative. Other important agents ofdisturbance in terrestrial avian habitats include drought, hurricanes, and herbivory

owing to insect outbreaks or mammals (Rotenberry et al. 1995, R. A. Askins,personal communication).

GENERAL BACKGROUND

Habitat Ecology of Birds

The proximate cues used by birds to discriminate among habitats probably includefactors that influence the availability of food, risk of predation (to adults or nests),

and availability of nest sites. These factors include vegetation structure, floris-tic composition, densities of conspecifics and heterospecifics, and microclimate(Hild6n 1965, James 1971, Hutto 1985, Block & Breunan 1993, Martin 1998).Cues at the landscape or regional scale may include the size, shape, distribution,

configuration, and connectedness of different patch types (Wiens et al. 1993). Dis-turbance likeLy-affects all aspects of avian habitat quality and selection from themi6rohabitat to regional scale.

Conservation Status of Species in Disturbance-MediatedHabitats and Ecosystems

In a review of the conservation status of birds of eastern North America, Askins

(1993) reported a significantly greater proportion of recent declines in estimatedabundances for species most often associated with disturbance (grasslands,shrubland, or savannas) than with forest birds (also see Peterjohn & Saner 1999).Estimates of trends in abundances were derived from the North American Breeding

Bird Survey (BBS), a continental avian survey program that was initiated in 1966


(Saner et al. 2000). We expanded the comparative approach of Askins (1993) toconsider all native North American terrestrial species for which abundance trendsfrom 1966 to 1998 could be estimated with reasonable precision. We included 274

species and placed each into one of five habitat categories: grassland (n = 27),shrub-scrub (n = 79), open woodlands and savanna (n = 63), closed-canopyforest (deciduous, coniferous, or mixed, n = 77), and generalist (n = 28). The

last category was used for species associated with two or more categories over allor part of their range and species of urban or suburban habitats. We then consultedSauer et al. (2000) for estimates of population trends. The significance of these esti-mated trends are placed by Saner et al. (2000) into four categories: significant (p <0.10), upward, or downward, and insignificant (p > 0.10), upward, or downward.

The distribution of species in the four trend categories varied significantly

among the different habitats or ecosystems (p = .0004, exact test of proportions). Agreater proportion of species experienced significant decreases in all disturbance-mediated habitats than in the forested or generalist categories (Figure 1). The

greatest proportion of decreases was within grassland birds where 56% of speciesin our sample declined significantly, followed by shrub-scrub (39%), and openwoodlands (33%). Combining these three categories into a general "disturbance"category revealed that 40% (n = 169) of North American species associated with

m

n

_,,45O

30

0Grassland Shrub-Scrub Savanna Forest Generallst

Habitat Groups

_'_ Signn. _ Decremm _ Sign.. _ IncreaseDecresse Increase

Figure 1 Status of North American birds in five habitat groups with respect to trends inabundance from 1966-1998. Trends from the North American Breeding Bird Survey (see

text for explanation of survey, how species were assigned to habitat groups, and how trendswere estimated and significance was determined).

DISTURBANCEECOLOGY AND BIRDS 255

some type of disturbance-mediated habitats were significantly decreasing between1966 and 1998. Conversely, only 17% of the disturbance species experiencedsignificant increases, whereas 34% of the forest species in our sample significantlyincreased.

EFFECTS OF FIRE

Under natural conditions, fire is the major force that governs the development andfunction of most deciduous, coniferous, and mixed forest types at temperate lati-tudes (Wright & Bailey 1982, Attiwill 1994, Frost 1998). Fire, along with otherfactors, is key in maintaining entire ecosystems such as grasslands and savannasat temperate and tropical latitudes, and Mediterranean-type shrublands (Moreno &Oechel 1994, McPherson 1997). In an extensive analysis of pre-Europeansettlement forests in the United States, Frost (1998) found that fire played afundamental role in the structure and composition of all but the wettest, mostarid, or fire-sheltered forest types. The estimated fire frequency from lightningstrikes and Native Americans varied from 1 to 700 y; one half of the continentalUnited States burned every 1 to 12 y. Other historical analyses of forests world-wide have also identified the pervasive role of fire on landscape diversity and thecomposition, structure, and function of stands (Atiwill 1994, Angelstam 1998).

Suppression of fire and longer intervals between fires have emerged over thepast 150 y in nearly all human dominated ecosystems and habitats where fire was afactor historically (Frost 1998, Smith 2000). Prescribed fire as a means of restoringor maintaining biodiversity is becoming commonplace throughout North Americaand elsewhere (Askins 2000), but long-term studies of periodic fire and birds arefew. We review four case histories: coniferous and boreal forests of the west-ern United States and southern Canada, oak savannas of the Midwestern United

States, pine savannas of the southeastern United States, and North American grass-lands. Other important systems in which fire has known effects on birds includeshrubsteppe in'.the Western United States (Knick & Rotenberry 1999), diverse

• Australian ecosystems (Reilly 1991), Mediterranen-type ecosystems (Moreno &

Oechel 1994), and Cerrado in Brazil (Cavalcanti 1999).?

Coniferous and Boreal Forests of the Western United States

and Southern Canada

The conifer-dominated forests of the western United States are variable in species

composition and include ecoregions (sensu Ricketts et al. 1999) that are geograph-ically widespread. The ecoregions include: Sierra Forests dominated by mixed-conifer stands (Ricketts et al. 1999); Rocky Mountain Forests (split by Rickettset al. 1999 into several regions) dominated at different elevations by ponderosapine (Pinus ponderosa), spruces (Picea spp.), lodgepole pine (Pinus contorta),and firs (Abies spp.); and mid-elevation forest in the southwest dominated by

256 BRAWN• ROBINSON• THOMPSON

ponderosapine. Fire was a common disturbance in these forests; estimates rangefrom 1per 10to 300y (Teller 2000),with stand replacementfires more common inmixed-conifer forests at relatively high elevations. Historically, replacement fireswere uncommon in open ponderosa pine forests (Telfer2000). Boreal forests aremost extensive in southern Canada, with elements found in northern sections ofthe continental United States.These forests--dominated by spruce,jack pine (Pi-nus banksiata), and quaking aspen (Populus tremuloides)--experience relativelyfrequent stand-replacing fires that occur every 40 to 100y (Telfer 2000).

The effects of fire on birds in these ecoregions are comparatively well known(Bock et al. 1978, Raphael et al. 1987). The most celebrated case history is thedependence of the Kirtland's Warbler (Dendroica kirtlandii) on burned jackpinehabitat (Probst & Weinrich 1993).Hutto (1995) examined effects of standreplac-ing fireson breeding birdcommunities inconiferous forests of theNorthern RockyMountain Forests. Fifteen species of birds were markedly more common on re-cently (i.e., within 1-3 y) burned habitats than in nearly all the other forest typesand successional stages sampled in the region. The Black-backed Woodpecker(Picoides arcticus)was largelyrestricted to the recently burned stands (1-2 y afterfire). Overall avian community structure was distinctive in burned sites (either1-2 y or 10-40 y after fire) largely because of greater densities of woodpeckers,flycatchers, or granivorous species. An open stand with many snags (i.e., dead ordying trees) and open ground provides for the nesting and foraging requirementsof these species. Other studies in mixed-coniferous forests confirm that cavity-nesting birds in particular benefit from the action of fire disturbance that createsnesting foraging substrate in snags (Caton 1998).

Information on the effects of wildfire and prescribed fireon birds in southwest-ern ponderosa pine forests, although hampered by a lack of long-term replicatedstudies (Finch et al. 1997), indicatesthat several species benefit from disturbance.And inmany cases, these are the same speciesor species groups that respond pos-itively to disturbance in mixed-conifer bums (summarized in Finch et al. 1997).Populations of woodpeckers, certain flycatchers, and certainground foragers tendto increase after fire.

Boreal forests in southern Canada have provided opporttmities to assess the ef-fects of natura$burn cycles. As with mixed-conifer sites, stand-replacementburnshost"distinctivebird communities from the time of initial bums and successionfrom brush/shrub to relatively mature forest (Helle & Neimi 1996, Imbeau et al.1999,Drapeau et al. 2000, Schieck & Hobson 2000). Virtually all studies have ob-served enhanced/_ diversity with fire disturbance and greater landscape diversity.The Black-backedWoodpecker,inparticular, is a specialistin recent burns (Book &Book 1974).This species forages onwood-boring beetles that inhabitdyingconifersfor 2-3 y after exposure to fire (Murphy & Lehnhausen 1998).

Midwestern Oak Savannas and Woodlands

Oak savannas andwoodlandswere formerlya significantelementof the landscapesof the midwesternUnitedStates and Canada that coveredan estimated11x 106 ha


(Nuzzo 1986, Anderson & Bowles 1999; important oak savannas are also foundin the southwest United States and California, but are not considered here). Oak

savannas and woodlands are distinguished from oak forests by a more open canopy(10% to 70% closure) and a herbaceous understory. Dominant tree species varyaccording to soil type and geographic location but typically include fire-resistantspecies such as bur oak (Quercus macrocarpa), blackjack oak (Q. marilandica),pin oak (Q. palustris), and white oak (Q. alba) (McPherson 1997). Under naturalconditions, the understory is dominated by herbaceous plants that are typically

associated with prairies (McPherson 1997). Oak savannas and woodlands wereformerly transitional vegetation types between prairies and closed-canopy forests(Anderson & Bowles 1999). Fire was the primary disturbance force in the region,and savannas burned less frequently than prairies, but more often than areas offorest.

Because of development and lack of disturbance, only 1% of the former oaksavannas in the Midwest remain, and this habitat type is judged as globally signif-

icant (Ricketts et al. 1999). With fire suppression, shaded conditions develop, theherbaceous understory is lost, and encroachment by mesophytic species such asmaples (Acer spp.) occurs (Packard 1993). Accordingly, prescribed fire, mechani-cal cutting, and herbiciding for the control of exotics are now common throughoutthe region in an effort to restore degraded savannas or maintain an open canopy(Packard 1993). Extant savannas in the upper Midwest are characteristically small,isolated renmants (Temple 1998).

Studies of birds and oak savannas have emphasized within-patch effects of

prescribed burning. Overall, it is clear that bird communities of oak savanna reflectthe transitional status of this vegetation type. Temple (1998) concluded that most

bird species characteristic of oak savannas also have affinities with either prairiesor forests. In the midwestem United States, this group includes about 40 bird

species; of these, about 60% are decreasing significantly in the region (Brawn1994). Whereas Temple (1998) concluded that there are no true savanna birds in theMidwest, one possible exception may be the Red-headed Woodpecker (Melanerpes

erythocephalus)---a species that is decreasing throughout much of its range (Saueret al. 2000). In Illinois' upland habitats, this species is restricted to oak savannasand habitats that mimic savannas such as parks (J. D. Brawn, unpublished data).

,:Bird communities in oak forest in the upper Midwest change with fire andsavanna restoration. Davis et al. (2000) reported that sites with a history of more

fires have fewer trees, higher densities of snags, and variable densities of shrubs.

Bird species richness over two breeding seasons was markedly greater in the burnedsites, and avian community structure varied systematically with the intensity ofrestoration. Canopy foliage gleaners such as the Red-eyed Vireo (Vireo olivaceus)and the Scarlet Tanager (Piranga olicavea) decreased with restoration, whereas

omnivorous ground or shrub feeders and woodpeckers increased. Species favoredby restoration and disturbance included the Baltimore Oriole (lcterus galbula), theRed-headed Woodpecker, and the Field Sparrow (Spizella pusilla). Many speciessuch as the Indigo Bunting (Passerina cyanea), Field Sparrow, and AmericanGoldfinch (Cardeulis tristis) breed regularly in oak savannas and nest or forage in


shrubs (Sample & Mossman 1994, Davis et al. 2000) and are also characteristic ofedge habitats and successional habitats such as abandoned pastures. Frequent andpervasive bums will eliminate shrubs; therefore, the frequency of fire disturbancecan significantly change the diversity of bird communities within restored oaksavannas and woodlands.

Southeastern Pine Savannas

The upland landscapes of the southeastern United States were once dominated byopen pine savannas with an herbaceous understory (Platt 1999). These savannascovered an estimated 200 x 106ha (Platt 1999) and were dominated by longleaf

pine (Pinuspalustris) (McPherson 1997). At elevations below 75 m, these savannaswere nearly continuous along the Atlantic and Gulf coasts from North Carolina toTexas.

Frost (1998) estimated that lightning caused fires every 1-3 y over much of the

region. The pine savannas of the southeastern United States, therefore, are one ofthe most pyrogenic vegetation types known. Not surprisingly, changing land use,fire suppression, commercial logging, and grazing have had immense impact onthe vegetation of the region. By 1985, only 4% of the pre-European settlementsavannas remained (Noss 1989, Platt 1999) with only 3000 ha in old-growth form.

The avifauna of the region includes three species that are essentially endemic;the Red-cockaded Woodpecker (Picoides borealis), the Brown-headed Nuthatch

(Sittapusilla), and the Bachman's Sparrow (Aimophila aestivalis) (Jackson 1988).The range and abundances of the Red-cockaded Woodpecker have decreaseddramatically; this species is classified as federally endangered. The Bachman's

Sparrow has also decreased significantly throughout much of its range, althoughit can use early successional pine fields (Dunning & Watts 1990). Fire-disturbedhabitat characterized by large, well-spaced pines and lack of midstory trees

(especially encroaching hardwood species) is essential for the Red-cockadedWoodpecker (James et al. 1997, Plentovich et al. 1998). Habitat requirements forthe Bachman's Sparrow can vary geographically, but this species was historically

also dependent on fire (Dunning & Watts 1990, Plentovich et al. 1998).The significant historical role of fire in molding the upland landscapes of the

southeast, coupled with the endangered status of Red-cockaded Woodpeckers

has generated considerable interest in the study of fire and avian biodiversityin the region (Wilson et al. 1995). A 15-y study of fire suppression in a Floridapine woodland documented progressive changes in habitat structure and floris-tics as the herbaceous understory was lost and mesophytic hardwoods encroached

(Engstrom et al. 1984). Avian species composition changed concomitantly, withopen-habitat species such as the Bachman's Sparrow and Loggerhead Shrike (La-nius ludovicianus) disappearing within 9 y as species characteristic of hardwoodforest colonized the site.

Habitat management for Red-cockaded Woodpeckers, such as short intervalsbetween growing season fires and removal of midstory, has provided opportunities


to assess how disturbance influences overall community structure. These studies

provide evidence that management for an obligate disturbance species can haveimportant collateral benefits for the regional avifauna. Burger et al. (1998) re-ported that intense management for woodpeckers promotes local increases in sevenspecies that are declining regionally or nationally. These species occur regularlyin early successional habitat or mature pine-grassland habitat with an herbaceousunderstory. Management units for Red-cockaded Woodpeckers in South Carolinaalso attracted over 30 scrub and successional species soon after burning and thin-

ning of the canopy and midstory (Krementz & Christie 1999).

Grasslands

Before European settlement, about 162 × 106 ha of grasslands occurred in NorthAmerica in three major ecoregions: tallgrass prairie, shortgrass prairie, and mixed-grass prairie (Samson & Knopf 1996, Herkert & Knopf 1998). Loss of nativegrassland habitat has been nearly complete in all regions (Samson & Knopf 1996),resulting from conversion to cropland, planting of exotics, overgrazing, and en-croachment by shrubs (Vickery et al. 1999). The biogeographic history of NorthAmerican grassland is complex, but these ecosystems undoubtedly developed withfrequent fire from lightning strikes and fires set by Native Americans (see severalchapters in Samson & Knopf 1996). The frequency of fire and other disturbanceagents such as grazing or mowing (a detrimental form ofanthropogenic disturbancedepending on when it occurs; see Bollinger et al. 1990) is a major determinant ofthe structure and plant species composition of most types of prairie. Vickery et al.(1999) identified 32 bird species that are obligate to temperate-zone grasslandsin North America, and they listed an additional 52 species that regularly occur in

grasslands or ecotonal grasslands such as savannas.Managing disturbance is the foundation of grassland restoration and conserva-

tion (Leopold 1949). Surprisingly few studies have been designed to assess fireeffects on grassland birds, but it is evident that the temporal and spatial scalesof disturbance have important consequences for habitat suitability. A 23-y study

by Johnson (1997) in mixed-grass prairie in North Dakota assessed the effects ofa 3-5 y burn rotation on avian abundances. Burning initially rid sites of vegeta-tioh and created open ground or short herbaceous vegetation; succession then led

to taller vegetation and encroachment of woody vegetation. Avian responses tofire disturbance and subsequent succession fall into three groups: 1. those speciesthat immediately colonized such as the Killdeer (Charadrius vociferous), Marbled

Godwit (Limosafedoa), or Upland Sandpiper (Bartramia longicauda); 2. speciesthat used sites about 2 y after burning, but before woody encroachment, such as theBobolink (Dolichonyx oryzivorus) and Grasshopper Sparrow (Ammodramus sa-vannarum); and 3. species that required woody vegetation and relatively long-termprotection from disturbance such as Common Yellowthroats (Geothlypis trichas).Similar patterns of turnover were found by Madden et al. (1999) in mixed-grassprairies in North Dakota that bum frequently. Therefore, even within the constraints


of what is a prairie or grassland, there is an important short-term successional cy-cle that is disturbance-dependent, with significant turnover in avian community

structure. Fire is particularly important in mesic grasslands that are comparativelyproductive and where woody encroachment proceeds rapidly (Madden et al. 1999).

Time elapsed since fire disturbance also influences bird populations in tallgrassprairie. Local abundances of Henslow's Sparrows (Ammodramus henslowii), aspecies of conservation concern throughout much of its range (Peterjohn & Saner1999), are highly sensitive to near-term fire history. In Illinois, newly burned sitesare generally avoided and greatest abundances are attained on sites 2--4 y postburn(Herkert & Glass 1999) within a relatively specific range of vegetation heights andlitter depths.

EFFECTS OF FLOODPLAIN DISTURBANCE

Floodplains are among the most disturbance-prone ecosystems. Although floodpulses tend to be predictably seasonal in most regions of the earth, the severity andtiming of the flood pulse vary greatly (Malanson 1993, Sparks 1995). Floods createoxbow lakes and backwaters, and promote primary successional gradients as plantcommunities form on newly created soils inside meander loops (point bars) andon islands (Malanson 1993). By depositing silt on the floodplain, floods enhance

primary productivity (Johnson et al. 1994). Severe floods can scour most or allvegetation from a floodplain; lack of severe flooding can allow woody encroach-ment on low-lying sections (Johnson 1994). The "flood pulse concept" (Junk et al.1989) recognizes the importance of seasonal floods in promoting high biodiversityand productivity in the lateral floodplain. The connection between rivers and theirbackwaters is crucial for wildlife, especially fishes (Sparks 1995) and waterfowl

(Bellrose et al. 1983). The intermediate disturbance hypothesis (Counell 1978)has been implicitly recognized in studies of floodplain ecosystems (Conner & Day1976). Floodplain disturbances create an extraordinary diversity of habitats in river

systems where fl'oods are not constrained or altered and terrestrial bird communi-

ties of floodplains are some of the richest in the world (e.g., Terborgh et al. 1990).The.river contfiaiaum concept (Varmote et al. 1980) relates the magnitude and pre-dictability of seasonal flood pulses to the size of the river or stream with maximaldiversity predicted in midsized rivers where levels of disturbance are intermediate.

Perhaps because of their high productivity and rich soils, floodplain habitatsare also among the most endangered ecosystems (Dynesius & Nilsson 1994).Rivers have been dammed, channelized, straightjacketed by levees, subjected

to increasing agricultural runoff, and lost to agriculture and plantation forestry(Hunt 1988, Sparks 1995). Floodplains have therefore lost much of their fish andwildlife as once-productive backwater swamps, marshes, and lakes have been iso-lated or subjected to excessive siltation when they do flood (Bellrose et al. 1983,Hunt 1988, Sparks 1995). In North America, three of the four terrestrial bird

species that became extinct since European settlement depended largely on flood-plain forests, including the Ivory-billed Woodpecker (Campephilus principalis),


Carolina Parakeet (Conuropsis carolinensis), and Bachman's Warbler (Vermivorabachmanii) (Askins 2000). Riparian forest loss in western North America is usuallyestimated at 95% or greater as a result of conversion to agriculture and alterationof hydrology (Knopf et al. 1988).

In this section, we review the linkages between disturbances and floodplainbird communities in three regions in North America: desert rivers of the southwestUnited States, big rivers of the central United States, and forested rivers of thesoutheastern United States.

Riparian Forests of Southwestern North America

Western riparian forests provide important resources for birds (Griunell & Miller1944, Rice et al. 1984, Rosenberg et al. 1991). A typical southwestern riparian

successional gradient consists of sandbars in the frequently flooded river channel,willows (Salix spp.) along the edge of the channel and on islands, cottonwoods(Populus spp.) farther back from the river channel, and, in areas less frequentlyflooded, more diverse hardwoods. Each plant community occurs in a well-defined,often narrow, zone with abrupt borders reflecting the frequency of disturbance(Rosenberg et al. 1991). These zones contain rich breeding bird communitiesthat may be associated with particular plant species for nesting and feeding.Disturbance-dependent willows, for example, are strongly preferred by severalspecies and subspecies that are currently rare and endangered. Examples are South-western Willow Flycatchers (Empidonax traillii extimus), Arizona Bell's Vireos(Vireo bellii arizonae), and Least Bell's Vireos (V.b. pusillus) (Askins 2000).

Alteration of the disturbance regimes in western riparian forests has threatened

this entire system (Carothers & Johnson 1975, Szaro & DeBano 1985, Knopfet al.1988, Ohmart 1994). Damming of rivers has reduced the extent of the willow-

cottonwood plant communities by reducing the severity and frequency of flooddisturbance (Rice et al. 1983, Anderson et al. 1983), with the resulting decline of

the many bird species that depend upon these communities (Strong & Bock 1990).Dams reduce seasonal flooding in undammed portions of the river, which driesout backwaters' and oxbows. Willows and cottonwoods do not grow along newly

created resery-oirs because water levels fluctuate too greatly to allow seed germi-nation (Hunter et al. 1987). Invasive introduced salt cedars (Tamarix chinensis)have taken over extensive areas of the floodplain in Arizona because dammingand irrigation projects have made the soil more salty and hence more favorable

to this species (Rosenberg et al. 1991). These dense homogeneous stands of saltcedar have far fewer birds than do native woodlands (Anderson et al. 1977). After

much trial and error, some of these riparian zones have been restored successfully(reviewed in Askins 2000), and flood pulses are being reintroduced to the ColoradoRiver in the Grand Canyon.

Large Rivers of Central North America

Large floodplain systems such as those found in the Midwestern United States werehistorically among the Earth's most productive ecosystems (Bellrose et al. 1983,

262 BRAWN • ROBINSON• THOMPSON

Sparks 1995), largely as a result of seasonal flood pulses that enriched soils andcreated diverse landforms and plant communities including backwaters, oxbowlakes, willow/cottonwood forests, mixed hardwood forests on natural levees, and

more open sandbar habitats. In the Midwest, these floodplains supported hugepopulations of migratory waterfowl (Bellrose et al. 1983) and diverse communitiesof terrestrial birds (Zimmerman & Tatschl 1975, Knutson et al. 1996).

Disturbance regimes within these systems have been altered by navigation dams,levees, and agriculture (Bellrose et al. 1983, Dynesius & Nilsson 1994, Johnson

et al. 1994). Most backwaters are isolated hydrologically or they rapidly fill with silt(Bellrose et al. 1983, Sparks 1995) during the summer. Levees increase the severityof floods in the river channel (Johnson et al. 1994), and flood-intolerant trees are lostin the remaining floodplain (Hosner 1958, Hunt 1988, Knutson & Klaas 1998).

As a result, remaining mixed floodplain forests are being replaced by skeletalforests of dead trees and willow thickets (Yin & Nelson 1995, Knutson & Klaas

1997). Before they fall, dead trees temporarily provide habitats for cavity-nestingbirds such as Red-headed Woodpeckers and Prothonotary Warblers (Protonotariacitrea) (Knutson & Klaas 1997).

With reduced flood severity, many rivers in the Great Plains that formerly lackedwoody cover are lined with mixed hardwoods more characteristic of floodplainseast of the Mississippi (Knopf & Scott 1990, Johnson 1994). Eastern forest birdsuse these new wooded corridors to invade western forests, activity that threa-tens the natural integrity of regional bird communities (Knopf et al. 1988). The lossof sandbar habitat resulting from altered disturbance regimes has been implicatedin the declines of the endangered Interior Least Tern (Sterna antillarum) and

Piping Plover (Charadrius melodus) (Schwalbach et al. 1993), which nest mostsuccessfully on island sandbars to which predators have little access (Kirsch 1996).

Southeastern North American Floodplain Forests

Southeastern floodplain forests have rich bird communities (Fredrickson 1980,Knutson et al. 1996, Sallabanks et al. 2000) that appear to depend in part on vary-ing levels of disturbance. Carolina Parakeets and Ivory-billed Woodpeckers mayhavo dependeffupon mature floodplain forest, a habitat that has been almost com-pletely logged and converted to rowcrops (Askins 2000). At least two species, theextinct Bachman's Warbler and the rare Swainson's Warblers (Limnothlypis swain-

sonii) historically inhabited dense stands of cane (Arundinaria gigantea) (Meanley1966, Eddleman et al. 1980, Hamel 1986, Remsen 1986), a plant that thrives atan intermediate level of floodplain disturbance and occasional fire. Human activ-ities have altered disturbance regimes to the extent that cane stands are now rare

(Ask.ins 2000). Several bird species depend largely upon tree species associatedwith varying levels of disturbance (Gabbe et al. 2001). Yellow-throated Warblers(Dendroica dominica), for example, depend upon bald cypress (Taxodium dis-tichum) growing in backwaters (Gabbe et al. 2001). There is little evidence, how-ever, that bird communities in this region change dramatically along successionalgradients as has been documented in rivers with unaltered floodplain dynamics


in, for example, Amazonia (Dyrcz 1990, Terborgh et al. 1990, Robinson 1997,Robinson & Terborgh 1997).

Riparian disturbances in southeastern floodplains create natural edges such asthe borders of oxbow lakes, natural levees, and the rivers themselves (Hupp &Osterkamp 1985). The natural edges may enhance both bird community diversityand nesting success (Sallabanks et al. 2000, Knutson et al. 2000). The complexvegetation structure along natural disturbances may provide dense cover in whichnests may be concealed (Knutson et al. 2000).

EFFECTS OF SILVICULTURAL DISTURBANCE

Nearly one half (1563 million ha) of the area of natural forests worldwide is con-sidered available for timber harvest given current legal and economic restrictions

(FAO 1999). In 1991, total world wood production was 4410 million m3, 37% ofwhich was in the tropics and the balance mostly in northern temperate and borealconifer forests (Whitmore 1998). Timber harvest is likely a more extensive distur-

bance agent in developed countries than all forms of natural disturbance. From theearly 1980s to the early 1990s, for example, approximately 24% of the timberlandin Michigan, Wisconsin, and Minnesota was disturbed by forest management (tim-ber stand improvement, harvesting, planting, etc.) and 13% by natural processes(fire, windthrow, flooding, and pests) (Schmidt et al. 1999).

Silvicultural treatments are typically applied to stands or patches with sim-

ilar tree composition and structure, often on a definable ecological land type.Regeneration treatments remove existing trees and establish tree reproduction.The clear-cut, shelterwood, and seed-tree methods regenerate even-aged standsand selection methods regenerate uneven-aged stands (Nyland 1996). Intermedi-ate treatments modify existing forest stands (stand improvement) or control its

growth and provide early financial returns (thinnings) (Nyland 1996).Uneven-aged and even-aged methods differ in the scale and intensity of distur-

bance. Uneven-aged methods maintain a mix of tree sizes or ages within a habitat

patch by periodically harvesting individual or small groups of trees. Even-agedmethods harvest most or all of the overstory and create a fairly uniform habitatpa_ch dominated by trees of the same age (Nyland 1996). At a landscape level,however, both methods can remove a similar volume of trees and regenerate similarareas of forest but the size and distribution of disturbance patches varies (Shirley

et al. 2000). Bird communities differ in their response to regeneration methods asa result of these differences in scale and intensity of disturbance.

For brevity, we review mainly the effects of regeneration practices in naturalor semi-natural North American forests where conservation is at least a secondary

goal, and do not consider intermediate treatments, intensive culture, or agroforestry.

North American Forests, Within Patch Effects

EVEN-AGEDMETHODS Even-age regeneration usually leads to near complete

turnover in the breeding bird community (Webb et al. 1977, Conner et al. 1979,


Franzreb & Ohmart 1978). About 50% of neotropical migratory birds in centralUS hardwood forests prefer disturbance-mediated early successional stands cre-ated by harvest and 10-20% prefer these habitats in southern pine forests (Dicksonet al. 1995).

In central and northern hardwood forests in the United States, nearly all (n > 50)

species that previously bred in the mature forest abandon regeneration cuts(Annand & Thompson 1997). Bird communities respond significantly to the dis-turbance and ensuing successional changes. If residual live and dead trees areleft in cut-over stands, some species such as Scarlet and Summer Tanager (Pi-

ranga rubra), Great Crested Flycatcher (Myiarchus crinitus), and Red-belliedWoodpecker (Melanerpes carolinus) that use the canopy or tree boles may continueto use the stand. Species that soon colonize and use residual snags, trees, slashpiles,or a rapidly developing herbaceous layer include Eastern Bluebirds (Sialia sialis),Northern Flicker (Colaptes auratus), Winter Wren (Troglodytes troglodytes),Carolina Wren (Thryothorus ludovicianus), Field Sparrow, and AmericanGoldfinch. By the second growing season, developing ground vegetation, lowshrubs, and young trees attract numerous species, including Common Yellowthroat,Chestnut-sided Warbler (Dendroica pensylvanica), Mourning Warbler (Oporor-

nisphiladelphia), Swainson's Thrush (Catharus ustulatus), Yellow-breasted Chat(lcteria virens), Prairie Warbler (Dendroica discolor), and Blue-winged Warbler

(Vermivorapinus). By 10 y after timber harvest, tree reproduction forms a closedcanopy, most early successional species abandon the site, and mature forest speciesbegin to colonize (DeGraaf & Chadwick 1987, Thompson et al. 1996). Overall,densities of birds in young regenerating forests often are similar to or much greaterthan those in mature or midsuccessional pole-sized forests (Conner et al. 1979,Yahner 1986, Probst et al. 1992, Hagan et al. 1997, Hobson & Schieck 1999).Moreover, species richness may be greater in regenerating stands in the easternUnited States (Conner et al. 1979, Yalmer 1986, King et al. 2001).

Regeneration methods influence the future vegetation composition of a stand.For example, use of shelterwood or selection methods as opposed to clear-cuttingwill favor more _shade-tolerant species over intolerant species, such as western

hemlock over Douglas fir (Pseudotsuga menziesii) in western conifer forests, orsugar maple (,4cer saccharum) and American beech (Fagus grandifolia) over oaksin e_tern deciduous forests. The use of the seed tree method can directly control

the seed source for regeneration and is often used in southern pine forests. Speciescomposition is sometimes directly manipulated by the planting of improved treestock and often occurs with Douglas fir in western forests and loblolly (Pinus

taeda) and shortleaf pine (Pinus echinata) in southern forests. These shifts infloristic composition can alter an avian community because many bird species havedemonstrated preferences for various tree species as foraging substrates (Franzreb& Ohinart 1978, Holmes & Robinson 1981).

A few studies have reported reproductive rates for songbirds in forest patches

managed by even-aged management. In extensively forested northern NewEngland, predation rates on artificial ground and shrub nests were not different


among successional stages (DeGraaf & Angelstam 1993). Elsewhere, Morse &Robinson (1999) observed lesser nest predation of Kentucky Warblers in matureforest than in recent clear-cuts. Aunand & Thompson (1997) reported compara-tively high nest success and low parasitism rates for several shrub nesting birdsin stands regenerated by the clear-cut or shelterwood methods in Midwestern UShardwood forests.

Even-aged management regenerates an entire stand, so stand size determinesthe size of the new habitat patch. Krementz & Christie (1999) found no evidence,however, that species richness or reproductive success varied among stands grow-

ing after clear-cuts ranging from 2.8 to 56.7 ha in southeastern longleaf pineforests.

UNEVEN-AGEDMETHODS Single-tree and group-selection regeneration methodsremove only a portion of the trees in a habitat patch and attempt to maintain aconstant tree size/class structure over time. As a result, the turnover in bird speciesfollowing a regeneration harvest, and during succession between harvests, is less

significant than that following even-aged regeneration harvests (Thompson et al.1995).

In eastern deciduous forests, most mature forest species occur in similar or

slightly lower numbers in forests managed by the selection method than in un-treated mature forest (Annand & Thompson 1997, Robinson & Robinson 1999).

Gap-dependent species such as Hooded Warblers (Wilsonia citrina), IndigoBunting, White-eyed Vireo (lqreo griseus), and Carolina Wren, or species de-pendent on understory density such as the Worm-eating Warbler (Helmitherosvermivorus), are more abundant in forests managed by the selection methodthan in mature even-aged forests. In New England's northern hardwood forests,bird communities in uneven-aged forest managed by selection method are alsosimilar to those in mature forests (DeGraaf & Chadwick 1987). Notwithstand-

ing, there is some evidence that uneven-aged management may selectively fa-vor red maple (Acer rubrum), a change that could result in less oak and beechmast for birds and other wildlife (R. A. Askins, personal communication). Se-

lection cuttingalso appears to have minimal impact on the composition andabundance of bird communities in western conifer forest. Similarly, in mixed-cotiifer forest in the western Sierra Nevada, there were few consistent differencesin bird communities between uneven-aged and mature even-aged forests (Morrison

1992).Again, there is limited knowledge of reproductive success in forest patches

managed by uneven-aged methods. In forests managed by the selection methodin southern Illinois, reproductive success of four songbird species was not, onaverage, greater in logged or nonlogged forest, and cowbird parasitism was onlygreater for two species in recent cuts (Robinson & Robinson 2001). Nest suc-cess of early succession birds does not appear to differ between group selectioncuts and clear-cuts in eastern deciduous forests (King et al. 2001, Clawson et al.2001).


North American Forests, Landscape Level Effects

Landscapes managed by even-aged methods are often a mosaic of different-agedstands, whereas those managed by uneven-aged methods are often less heteroge-neous and composed of a range of tree sizes. Thompson et al. (1996) used a simplemodel to demonstrate how these patterns in composition might affect/_ diversity incentral hardwood forests. The biggest differences among landscapes managed bythese silvicultural methods was not in the abundance of species that nest in mature

forest, but in early succession species that favored either small gaps (created by se-lection cutting) or large patches (created by clear-cutting). Thompson et al. (1992)compared the abundance species in landscapes with and without even-aged man-agement (clear-cutting). In general, the abundance of only a few species nesting inmid- or late-successional forest differed significantly between landscapes, and/_

diversity and abundance of early successional species was greater in managed land-scapes. These patterns of diversity vary with rotation time between harvests, but sil-vicultural disturbance clearly has a potentially beneficial effect on avian diversity.

Nonetheless, timber harvest decreases the continuity of mature or old forestin the landscape and is a source of habitat fragmentation. Forest fragmentation

by human dominated land-uses, such as agriculture or urban and suburban de-velopment, is hypothesized to reduce reproductive success of forest songbirds byelevating numbers of generalist predators and nest predation in the landscape or

by increasing the amount of edge and edge-related nest predation (Robinson et al.1995). In one of the few controlled landscape-scale studies of the effects of regen-eration practices, Clawson et al. (2001) estimated differences in abundance and

reproductive success of common forest birds in Midwestern oak forests before andafter logging. Densities of some forest species decreased while others increasedfollowing logging; however, neither nest predation or cowbird brood parasitismincreased following logging.

Syntheses are hampered by differences in methods (especially artificial vs.real nests), the potential effects of other uncontrolled local factors, or effects oflandscape context (i.e., Donovan et al. 1997). Manolois (2001) reviewed evidencefor edge effects resulting from timber harvest and concluded that many of the

" studies that rep0rted no edge effects did not have adequate statistical design orpower to detect effects if they were present, and 9 of 13 studies with adequatepower detected edge effects. General pooling of species may be misleading sincelower nesting success near edges may be specific to certain groups such as groundnesters (Flaspohler et al. 2001). Overall, evidence is conflicting for the nature ofedge effects between mature forest and recently harvested stands (Yalmer & Scott1988, Morse & Robinson 1999, Manolis et al. 2001).

Conservation Implications and Research Needs

This paper and other analyses (e.g., Alverson et al. 1994, Askins 2000) summa-rize evidence that ecological disturbance is fundamental to the conservation ofbirds. Whereas outright habitat loss is the most direct threat to avian biodiversity


worldwide, it is clear that significant numbers of bird species are associated withhabitats and ecosystems created by diverse forms of disturbance. We do not min-imize the conservation significance of species associated with habitats that arerarely subjected to large-scale disturbance (e.g., primary or old growth forests),and we are encouraged that forest birds have benefited by afforestation in theeastern United States (Askins 2000). Rather, we believe that understanding and

applying ecological disturbance offers opportunities to conserve large and diversegroup species--many of which are declining as a result of habitat loss and succes-sional changes in habitat structure.

The frequency and intensity of ecological disturbance is a determinant of thepresence of many terrestrial habitat types (or ecosystems) in North America thatsupport significant components of avian biodiversity (Figure 2). We believe thisoffers significant management and research opportunities; if conservation strate-gies for birds are to be effective, some combination of these habitats (along withtracts of mature forest) should be maintained within or across landscapes. Pro-

tected habitat should not be protected from processes that maintain or conservesignificant elements of biodiversity.

As shown here, the bulk of information about disturbance and birds is on varia-

tion in populations and communities resulting from the initial effects of disturbance

High Young, Even-Aged Shrub/Scrub GrasslandForest

8¢-

- i- Savanna

"6 Woodlands

_ . Matte, I-yen-Aged Mature,Uneven-Aged-- Forest Forest

_o"_ HighFrequency of Disturbance

Figure 2 Selected terrestrial habitats used by birds inNorth America qualitatively ordinatedaccording to frequency and intensity of disturbance. Even-aged forest can result from fire,wind throw, floods, or even-aged silviculture. Uneven-aged forest results from small-scalewind throw or uneven-aged silviculture. Woodlands and savanna result from moderate to lightintensity frequent fires. Grasslands generally result from frequent and intense fire, gazing,or both.


or ensuing succession. These studies are highly informative to land managers, butthe need for comprehensive understanding of disturbance and birds motivates acomplex research agenda. A prevailing need is to design and implement long-termand large-scale studies that monitor the effects of disturbance. Studies to date havegenerally been short-term, without controls, unreplicated, and unplanned (Finchet al. 1997). Outlined below are the issues and questions that we believe are themost pressing.

Associations Between Habitat Disturbance

and Population Viability

Studies of forest birds, habitat fragmentation, and edge effects clearly identifythe usefulness of information on annual reproductive success and other demo-

graphic traits that affect population viability (Heske et al. 2001). Data on localabundances are crucial but incomplete in the context of estimating viability and

assessing land-management options (Brawn & Robinson 1996). With few excep-tions (e.g., Johnson & Temple 1990, Krementz & Christie 1999), little is knownabout the demography of birds in disturbance-mediated or successional habitats.Our expectation is that fecundity will change with successional changes in habitat,especially in grasslands and early successional forests, but details on the time-frameof this dynamic will be highly informative to managers in determining desirableintervals between prescribed fire, silvicultural treatments, or other modes of dis-

turbance. Comparative studies from various geographic locations will be highlyinformative. In the Midwest, for example, prescribed fire and restoration maintainoak savannas in a specific disturbance-mediated state, but long-term successional

changes are common following the less frequent stand-replacing fires in coniferousforests.

The effects of patch size on disturbance-dependent birds also requires closestudy. Numerous studies confirm that avian diversity and reproductive successis, on average, greater in larger than smaller patches of forest and grasslands

(Thompson et al_2001, Heske et al. 2001). The effects of patch size on birds in moretransitional, disturbance-mediated habitats such as early successional forests andoak savannas are less clear. Early results from a study comparing nesting successof f6rest birds in managed forests found little or no difference in nesting successof disturbance-dependent birds in small selection openings vs. larger clear-cuts(Clawson et al. 2001; see also King et al. 2001).

Disturbance and Landscape Context

Questions about the scale and landscape context of disturbance will be differ-ent depending on the mode of disturbance and the ecoregion involved. In theextensive coniferous forests of the western United States and southern Canada,

a primary question is how to apply or allow disturbance so that a natural ornatural-like mosaic pattern of cover types is maintained (Hunter 1993, Landreset al. 1999). In isolated oak savanna restorations, the question is the effect of


surrounding habitat; that is, what are the effects (if any) of nearby grassland, for-est, or urban areas on the structure of savanna bird communities (Temple 1998)?Specific issues include the dispersion and arrangement of disturbance patches.Such studies will also serve to address questions from theory on ecological neigh-borhoods and the scale of ecological patterns or processes (Addicott et al. 1987,Engstrom & Mikushinski 1998). The frequency and extent of natal and breedingdispersal are key to determining the scale at which birds respond demographicallyto landscape features (Dunning et al. 1992, Vos et al. 2001) and are a persis-tent uncertainty for most species. Notable exceptions are the model studies onthe Bachman's Sparrow and the Red-cockaded Woodpecker that relate ecologi-cal disturbance, landscape features, and the spatial structure of their populations(Dunning et al. 1992, Engstrom & Mikusinski 1998, James et al. 1997). Ide-ally, local conservation decisions regarding ecological disturbance should be madewithin a multi-scale framework that acknowledges landscape and regional goalsbased on historical heterogeneity in ecological landtypes (Thompson & DeGraaf2001).

Effects of Natural vs. Anthropogenic Disturbance

Future land management planning will have to consider a range of approaches tobalance economic needs and conservation objectives. Understanding how these

different approaches affect the size and viability of birds associated with distur-bance is essential. For example, a stand-replacing fire and a clear-cut both removetrees but the long-term implications for avian conservation are likely to be different(I-Iobson & Schieck 1999, Imbeau et al. 1999). One approach to the conservationof disturbance-dependent birds is to restore natural disturbance regimes, but thisis realistic only on large expanses such as those in the western United States. Else-where, disturbance-dependent birds are likely to depend on human activities thatstem from direct management or commercial activities. Ways to optimize these ac-tivities for conservation, renewable resources, and agriculture requires immediate

study. In forest_ with commercial value, for example, the range of options includes(from most to least intensive): short rotation intensive culture, plantation forestry,commercial forestry in natural stands, multiple-use or ecosystem management,and wilderness management through natural disturbance. Another issue is the in-

tegrity of floodplain bird communities that depend upon restoration of flood-pulsedisturbances (Galat et al. 1998); at present, there is disagreement over whether theflood pulse should be mimicked through moist soil management (Galat et al. 1998)or promoted by directly allowing the river to flow into sections of the floodplainduring flood pulses.

Finally, we believe that the social implications of ecological disturbance requireexploration. Episodes of opposition to prescribed fire and silviculture, for example,strongly identify the need to educate the public on the positive role that naturaland anthropogenic disturbance can play in the conservation of avian biodiversity(Marynowski & Jacobson 1999).


ACKNOWLEDGMENTS

We thank Bob Askins, Dick DeGraff, Jim Herkert, and David King for thought-

ful comments. The help of Kathy Jennings was instrumental in preparing the

manuscript. Research support to JDB and SKR was provided in part by the

National Science Foundation (DEB 99-04058).

Visit the Annual Reviews home page at www.AnnualReviews.org

LITERATURE CITED

Addicott .IF, Aho JM, Antolin MF, Padilla Bellrose FC, Havera SP, Paveglio FL Jr, Stef-DK, Richardson JS, Soluk DA. 1987. Eco- feck DW. 1983. The Fate of Lakes in the llli-

logical neighborhoods: scaling environmen- nois River Valley. Ill. Nat. Hist. Sur_. BioLtal patterns. Oikos 49:340--46 Notes 119

AlversonWS, Kuhlman W, WallerDM. 1994. Block WM, Brennan LA. 1993. The habitat

Wild Forests, Conservation BiologyandPub- concept in ornithology: theory and applica-lic Policy. Washington, DC: Island. 300 pp. tions. Curt Ornithol. 11:35-91

Anderson B, Ohmart R, Rice J. 1983. Avian Bock CE, Book JB. 1974. On the geographi-and vegetation community structure and their cai ecology and evolution of the three-toedseasonalrelationshipsintheLowerColorado woodpeckers, Picoides tridactylus and PRiver Valley. Condor 82:392-405 arcticus. Am. Mid. Nat. 92:397-405

AndersonRC, BowlesML. 1999.Deep-soilsa- BoekCE, RaphaelM, BockJB. 1978. Chang-vannas and barrens of the Midwestern United ing avian community structure during earlyStates. See Anderson et al. 1999, pp. 155-70 post-fire succession in the Sierra Nevada.

Anderson RC, Fralish JS, Baskin JM, eds. Wilson Bull. 90:119-23

1999. Savannas, Barrens, and Rock Outcrop Bollinger EK, Boilinger PB, Gavin TA. 1990.Plant Communities of North America. Cam- Effects of hay-cropping on eastern popula-bridge, UK: Cambridge Univ. Press. 470 pp. dons of the Bobolink. Wild. Soc. Bull. ! 8:

Angelstam P. 1998. Maintaining and restoring 142-50biodiversityin European boreal forests by de- Botkin DB. 1990. Discordant Harmonies. Ox-veloping natural disturbance regimes. 3. Veg. ford, UK: Oxford Univ. Press. 241 pp.Sci. 9:593-602 Brawn JD. 1994. An overview of avian commu-

Annand EM, Thompson FR III. 1997. For- nities in North American savanna habitats:

est bird response to regeneration practices in current knowledge and conservation need.central .hardwood forests, d. Wildl. Manage. Proc. North Am. Conf. Barrens Savannas, pp.61:159-71 145-46. Chicago, IL: US EPA, Great Lakes-

Askins RA. 1993. Population trends in grass- Nat. Prog. Off.land, shrubland, and forest birds in eastern Brawn JD, Robinson SK. 1996. Source-sinkNorth America. Curt OrnithoL 11:1-34 dynamics may complicate interpretation of

Askins RA. 2000. Restoring North America "s long-term census data. Ecology 77:3--12Birds. Lessons from Landscape Ecology. Burger LW Jr, Hardy C, Bein J. 1998. EffectsNew Haven, CT: Yale Univ. Press. 288 pp. of prescribed fire and midstory removal on

Attiwill PM. 1994. The disturbance of forest breeding bird communities in mixed pine-

ecosystems: the ecological basis for con- hardwood ecosystems of Southern Missis-servative management. For EcoL Manage. sippi. See Pruden & Brennan 1998, pp. 107-63:247-300 13


landscapes: retrospective power analysis and Packard S. 1993. Restoring oak ecosystems.Minnesotaresults. Wildl. Soc. Bull. 28:1088- Rest. Manage. Notes 11:5-16I01 Peterjohn BG, Saner JR. 1999. Population sta-

Martin TE. 1998. Are habitat preferences of ms of birds of North American grasslandcoexisting species under selection and adap- birds from the North American Breedingfive? Ecology 79:656-70 Bird Survey, 1966-1999. Stud. Avian Biol.

Martin TE, Finch DM, eds. 1995. Ecology 19:27-44

and Management of Neotropical Migratory Petraitis PS, Latham RE, Nisenbaum RA.Birds. New York: Oxford Univ. Press 1989. The maintenance of species diver-

Marynowski SB, Jacobson SK. 1999. Ecosys- sity by disturbance. Q. Rev. Biol. 64(4):393-tern management education for public lands. 418Wildl. Soc. Bull. 27:134-145 Pickett STA, White PS, eds. 1985. The Ecology

Marzluff JM, Sallabanks R, eds. 1999. Avian of Natural Disturbance and Patch Dynamics.

Conservation: Research and Management. San Diego: Academic. 472 pp.Washington, DC: Island. 563 pp. Platt WJ. 1999. Southeastern pine savannas. See

McPherson GR. 1997. Ecology and Manage- Anderson et al. 1999, pp. 23-51ment of North American Savannas. Tucson: Plentovich S, Tucker JW Jr, Holler NR, HillUniv. Ariz. Press. 209 pp. GE. 1998. Enhancing Bachman's sparrow

Meanley B. 1966. Some observations on habi- habitat via management of red-cockadedtats of Swainson's Warbler. Living Bird woodpeckers. J. Wildl. Manage. 62(1):347-5:151--65 54

Moreno JM, Oechel WC, eds. 1994. The Role Probst JR, Rakstad DS, Rugg DJ. 1992. Breed-

of Fire in Mediterranean-Type Ecosystems. ing bird communities in regenerating and ma-Ecol. Stud. No. 107. New York: Springer- ture broadleafforests in the USA Lake States.Verlag. 201 pp. For. Ecol. Manage. 49:43--60

MorrisonML. 1992. Bird abundances in forests Probst JR, Weinrich J. 1993. Relating Kirt-

managed for timber and wildlife resources, land's Warbler populations to changing land-Biol. Conser_. 60:127-34 scape composition and structure. Land. Ecol.

Morse SF, Robinson SK. 1999. Nesting success 8:257-71

of a Neotropical Migrant in a multiple-use, Pruden TK, Brennan LA, eds. 1998. Fireforested landscape. Conser_. Biol. 13:327- in Ecosystem Management: Shifting the37 Paradigm from Suppression to Prescription.

Murphy EC, Lehnhausen WA. 1998. Density Tall 77mbers Fire Ecol. Conf. Proc. No. 20.and foraging ecology of woodpeckers fol- Tall Timbers Res. Stn., Tallahassee, FLlowing a stand-replacement fire. J. gqldl. Raphael MG, Morrison ML, Yoder-WilliamsManage. 62:1359-72 _ MP. 1987. Breeding bird populations during

Noss R. 1_189.Longleaf pine and wiregrass: twenty-fiveyearsofpostfiresuccessioninthekeystone components of an endangered Sierra Nevada. Condor89:614-26ecosystem. Nat. Areas J. 9:211-13 Reilly P. 1991. The effect of wildfire on bird

Nuzzo VA. 1986. Extent and status of Midwest populations in a Victorian Coastal Habitat.oak savanna: presettlement and 1985. Nat. Emu 91:100-6AreasJ. 6:6-36 Remsen JV Jr. 1986. Was Bachman's Warbler

Nyland RD. 1996. Silviculture: Concepts and a bamboo specialist? Auk 103:216-19Applications. St. Louis, MO: McGraw-Hill Rice J, Anderson BW, Ohmart RD. 1984. Com-

Ohrnart RD. 1994. The effects of human-in- parison of the importance of different habitatduced changes on the avifauna of western attributes to avian community organization.riparian habitats. Stud. Avian Biol. 15:273- J. Wildl. Manage. 48:895-91185 Ricketts TH, Dinerstein E, Olson DM, Loucks


Carothers SW, Johnson RR. 1975. Waterman- scapes and wildlife conservation. See De-

agement practices and their effects on Graaf&Miller 1996a, pp. 3-36nongame birds in range habitats. In Proc. Dickson JG, Thompson FRIII, Conner RN,Syrup. Manage. For Range Habitats Non- Franzreb KE. 1995. Silviculture in centralgame Birds, ed. DR Smith, pp. 210-22. and southeastern oak-pine forests. See Mar-Washington, DC: USDA For. Serv. GTR tin & Finch 1995, pp. 245-66WO- 1 Donovan TM, Jones PW, Annand EM, Thomp-

Caton EL. 1998. Postfire habitat use by cavity- son FR III. 1997. Variation in local-scale

nesting birds in Northwestern Manitoba. See edge effects: mechanisms and landscape con-Pruden & Brennan 1998, p. 364 (Abstr.) text. Ecology 78:2064-75

Cavalcanti RB. 1999. Bird species richness and Drapeau P, Leduc A, Giroux J-F, Savard J-PL,conservation in the Cerrado region of central Bergeron Y, Vickery WL. 2000. Landscape-Brazil. Stud. Avian Biol. 19:2A.A. a.9 scale disturbances and changes in bird com-

Clawson RC, Faaborg J, Gram WK, Porneluzi munities of boreal mixed-wood forests. Ecol.PA. 2001. Landscape level effects of forest Monogr. 70(3):423-44

management on bird species in the Ozarks of Dunning JB, Danielson BJ, Pulliam HR. 1992.southeastern Missouri. InProc. 2ndMissouri Ecological processes that affect populations

Ozark For Ecosystem Symp.: Post Treatment in complex landscapes. Oikos 65:169-75Results of the Landscape Experiment, ed. SR Dunning JB, WaRs BD. 1990. Regional dif-Shirley, JM Kabrick. USDA For. Se_ Gen. ferences in habitat occupancy by Bachman'sTech. Rep. St. Paul, MN. In press sparrow. Auk 107:463-72

Connell JH. 1978. Diversity in tropical rain- Dynesius M, Nilsson C. 1994. Fragmentationforests and coral reefs. Science 199:1302-10 and flow regulation of river systems in the

Conner RN, Via JW, Prather ID. 1979. Ef- northernthirdoftheworld.Science266:753-

fects of pine-oak clear-cuRing on winter and 62breeding birds in southwestern Virginia. 14ql- Dyrcz A. 1990. Understory bird assemblagesson Bull. 91:301-16 in various types of lowland tropical forest in

Conner WH, Day JW. 1976. Productivity and Tambopala Reserve, SE Peru (with faunistic

composition of a bald cypress-water tupelo notes). Acta ZooL Cracov. 33:215-33and bottomland hardwood site in a Louisiana Eddleman WR, Evans KE, Elder WH. 1980.

swamp. Am. J. Bot. 63:1354--64 Habitat characteristics and management ofDavis MA, Peterson DW, Reich PB, Crozier Swainson's Warbler in southern Illinois.

M, Query T, et al. 2000. Restoring savanna Wildl. Soc. Bull. 8:228-33using fire: impact on the breeding bird tom- Engstrom RT, Crawford RL, Baker WW. 1984.munity. Rest. EcoL 8:30-40 Breeding bird populations in relation to forest

DeGraaf RM, Ange}starn P. 1993. Effects of structure following fire exclusion: a 15-yeartimber _ize-class on predation of artificial study. Wilson Bull. 96(3):437-50nest in extensive forest. For. Ecol. Manage. Engstrom RT, Mikusinski G. 1998. Ecological61:127-36 neighborhoods in red-cockaded woodpecker

DeGraafRM, ChadwickNL. 1987. Forest type, populations. Auk 115:473-78

timber size class and New England breeding FAO. 1999. State of the World's Forests 1999.birds, d. Bqldl. Manage 51:212-17 Rome: Food Agric. Organ. UN

DeGraaf RM, Miller RI, eds. 1996a. Comer- Finch DM, Ganey JL, Yong W, Kimball RT,

vation of Faunal Diversity in Forested Land- Sallabanks R. 1997. Effects and interactionsscapes. London: Chapman & Hall. 633 pp. of fire, logging, and grazing. In Songbird

DeGraaf RM, Miller RI. 1996b. The impor- Ecology in Southwestern Ponderosa Pinetance of disturbance and land-use history in Forests: A Literature Review, ed. WM Block,

New England: implications for forested land- DM Finch, pp. 103-36. USDA For. Serv. Gen.


Tech. Rep. RM-GTR-292. Fort Collins, CO. Herkert JR, Glass WD. 1999. Henslow's spar-152 pp. row response to prescribed fire in an Illinois

Flaspohler DJ, Temple SA, Rosenfield RN. prairie remnant. Stud. Avian Biol. 19:160-2001. Species-specific edge effects on nest 64success and breeding bird density in a HeskeEJ, RobinsonSK, Brawn JD. 2001. Nestforested landscape. Ecol. Appl. 11:32--46 predation and neotropical migrant songbirds:

Franzreb KE, Ohmart RD. 1978. The effects piecing together the fragments. Wildl. Soc.

of timber harvesting on breeding birds in Bull. 29:52-61a mixed-coniferous forest. Condor 80:431- Hild6n O. 1965. Habitat selection in birds: a41 review. Ann. Zool. Fenn. 2:53-75

Fredrickson LH. 1980. Impact of water man- Hobson KA, Sehieck J. 1999. Changes in birdagement on the resources of lowland hard- communities in boreal mixedwood forest:wood forests. In Integrating Timber and harvest and wildfire effects over 30 years.Wildlife Management in Southern Forests, Ecol. Appl. 9:849--63ed. RH Chabreck, RH Hills, pp. 51--64. Holmes RT, Robinson SK. 1981. Tree speciesBaton Rouge: La. State Univ. preferences of foraging insectivorous birds

Frost CC. 1998. Presettlement fire frequency in northern hardwoods forest. Oecologiaregimes of the United States: a first approx- 48:31-35imation. See Pruden & Brennan 1998, pp. Hosner JF. 1958. The effects of complete inun-70-81 dation upon seedlings of six bottomland tree

Gabbe A, Robinson SK, Brawn JD. 2001. species. Ecology 39:371-73

Treespeciespreferencesofforaginginsectiv- HuntCE. 1988. Down By The River: The lmpactorous birds: implications for floodplain forest of Federal Water Projects and Policies On Bi-restoration. Conse_ Biol. In press oiogical Diversity. Washington, DC: Island.

Galat DL, Fredrickson LH, Humburg DD, 260 pp.Bataille KJ, Bodie JR, et al. 1998. Flood- Hunter ML Jr. 1993. Naturai fire regimes as spa-

ing to restore connectivity of regulated, large- tial models for managing boreal forests. Biol.river wetlands. BioScience 48:721-33 Conse_ 65:115--20

Grinnell J, Miller AH. 1944. The distribution Hunter WC, Anderson BW, Ohmart RD. 1987.

of the birds of California. Pacific Coast Avi- Avian community structure changes in a ma-fauna No. 27. Cooper Ornithol. Club, Berke- ture floodplain forest after extensive flood-ley, CA ing. J. Wildl. Manage. 51:495-502

HaganJM, McKinleyPS, MeehanAL, Grove Hupp CR, Osterkamp WR. 1985. Bottom-SL. 1997. Diversity and abundance of land- land vegetation distribution along Passagebirds in a northeastern industrial forest. J. Creek, Virginia, in relation to fluvial iand-

Wildl. Manage. 61:7_18-35 forms. Ecology 66:670-81Hamei PI_. 1986. Bachman's Warbler." A HuttoRL. 1985. Habitat selection by nonbreed-

Species in Peril. Washington, DC: Smithson. ing migratory land birds. In Habitat SelectionInst. in Birds, ed. ML Cody, pp. 455-76. Orlando,

Hanski I. 1991. Single-species metapopulation FL: Academic. 558 pp.

dynamics: concepts, models, and observa- Hutto RL. 1995. Composition of bird commu-tion. Biol. J. Linn. Soc. 42:17-38 nities following stand-replacement fires in

Helle P, Niemi GJ. 1996. Bird community Northern Rocky Mountain (U.S.A.) coniferdynamics in boreal forests. See DeGraaf & forests. Conse_ Biol. 9:1041-58Miller 1996a, pp. 209--34 Imbeau L, Savard JPL, Gagnon 1L 1999.

Herkert JR, Knopf FL. 1998. Research needs Comparing bird assemblages in successionalfor grassland bird conservation. See Marzluff black spruce stands originating from fire and& Sallabanks 1998, pp. 273-82 logging. Can. J. Zool. 77:1850--60


Jackson JA. 1988. The southeastern pine forest FB, Szaro RC. 1988. Conservation of ripar-

ecosystem and its birds: past, present, and ian ecosystems in the United States. Wilsonfuture. Bird Conserv. 3:119-59 Bull. 100:272-84

James FC. 1971. Ordinations of habitat rela- KnopfFL, Scott ML. 1990. Altered flows and

tionships among breeding birds. Wilson Bull. created landscapes in the Platte River head-83:215-36 waters, 1840--1990. In Management of Dy-

James FC, Hess CA, Kufrin D. 1997. Species- namic Ecosystems, ed. JM Sweeney, pp. 47-centered environmental analysis: indirect ef- 70. Lafayette, IN: The Wildl. Soc., Northfects of fire history on red-cockaded wood- Cent. Sect.

peckers. Ecol. Appl. 71:118-29 Knutson MG, Gutreuter SJ, Klaas EE. 2000.Johnson DH. 1997. Effect of fire on bird pop- Patterns of artificial nest depredation in a

ulations in mixed-grass prairie. In Ecology large floodplain forest. J. Wildl. Manage.and Conservation of Great Plains Verte- 64:576-83brates, Ecol. Stud. No. 125, ed.FLKnopf, FB KnutsonMG, HooverJP, KlaasEE. 1996. The

Samson, pp. 181-206. New York: Springer- importance of floodplain forests in the con-Verlag. 320 pp. servation and management ofneotropical mi-

Johnson KN, Sessions J, Franklin J, Gabriel J. gratory birds in the Midwest. See Thompson

1998. Integrating wildfire into strategic plan- 1996, pp. 168-88ning for Sierra Nevada forests. J. For. 96:42- Knutson MG, Klaas EE. 1997. Declines in49 abundance and species richness of birds fol-

Johnson RG, Temple SA. 1990. Nest preda- lowing a major flood on the Upper Missis-tion and brood parasitism of tallgrass prairie sippi River. Auk 114:367-80birds. J. Wildl. Manage. 54:106-11 Knutson MG, Klaas EE. 1998. Floodplain for-

Johnson RH, Sharitz RR, Dixon PM, Segal est loss and changes in forest communityDS, Schneider RL. 1994. Woody plant re- composition and structure in the Upper Mis-

generation in four floodplain forests. Ecol. sissippi River: a wildlife habitat at risk. Nat.Monogr. 64:45-84 Areas J. 18:138--50

JohnsonWC. 1994. Woodland expansion in the Krementz DG, Christie JS. 1999. Scrub-Platte River, Nebraska: patterns and causes, successional bird community dynamics inEcol. Monoge. 64:45-84 young and mature longleaf pine-wiregrass

Junk WJ, Bayley PB, Sparks RE. 1989. savannahs.J. Wildl. Manage. 63:803-14The flood pulse concept in river-floodplain Landres PB, Morgan P, Swanson FJ. 1999.

ecosystems. Can. Spec. Publ. Fish Aquat. Overview of the use of natural variabilityconcepts in managing ecological systems.Sci. 106:110-27

King DI, DeGraaf RM, Griffin CR. 2001. Ecol. Appl. 9:1175-88Productivity of early-successional shrubland Leopold A. 1949. A Sand County Almanac andbirds irzlzlear-cuts and groupcuts in an eastern Sketches Here and There. New York: Oxforddeciduous forest. J. Wildl. Manage. 65:345- Univ. Press. 226 pp.50 Madden EM, Hanson AJ, Murphy RK. 1999.

Kirsch EM. 1996. Habitat selection and pro- Influence of prescribed fire history on habi-

ductivity of Least Terns on the Lower tat and abundance of passerine birds inPlatte River, Nebraska. Wildl. Monogr. 60:1- northern mixed-grass prairie. Can. Field-48 Nat. 113:627--40

Knick ST, Rotenberry .IT. 1999. Spatial distri- Malanson GP. 1993. Riparian Landscapes.bution of breeding passerine bird habitats in New York: Cambridge Univ. Press. 296 pp.a shrubsteppe region of southwestern Idaho. Manolis JC, Anderson DE, Cuthbert FJ.Stud. Avian Biol. 19:104-11 2001. Patterns in clear-cut edge and frag-

Knopf FL, Johnson RR, Rich T, Samson mentation effect studies in northern-conifer

the role rbance in thecology and ofbirds* · silviculture stands apart as a commercially driven...

Documents