Northern Prairie Wildlife Research Center
Savannah sparrows were strictly migrants in Ohio during the nineteenth century. The first breeding populations were reported in several northern counties during the 1920s. Nesting pairs spread into southern and unglaciated counties during the 1970s, and Ohio is one of few states presently supporting an expanding breeding population (Robbing et al. 1986, Peterjohn and Rice 1991). Savannah sparrows are numerous in the northern and glaciated counties but rare and locally distributed in the southwestern and unglaciated counties (Peterjohn 1989).
The earliest migrants appear in Ohio during late March and early April. Nesting activity begins in late April. Most clutches are produced in May, the earliest published egg date being 10 May. Most first broods fledge between 15 June and 5 July. Renesting and second broods are frequent, producing clutches through mid-July and fledglings into early August. Fall migrants first appear in early September and reach maximum numbers between 25 September and 20 October. Savannah sparrows are rare and locally distributed winter residents in southern and central counties (Peterjohn 1989, Peterjohn and Rice 1991).
Savannah sparrow nests are placed on or near the ground within dense clumps of grasses. In Ohio, most breeding savannah sparrows are associated with grassy and clover hayfields, reclaimed surface mines, grassy fields bordering airports, and small grains. They prefer upland habitats, but are sometimes found within the grassy borders of lakes and marshes (Peterjohn 1989, Peterjohn and Rice 1991).
Savannah sparrows nest in a variety of open field habitats throughout North America, including arctic and subarctic tundra (Bent 1968), shrub-willow (Salix spp.) communities (Weatherhead 1979, Knopf et al.1988, Hendricks and Pidgeon 1990), and sandy beach dunes (Welsh 1975, Ross 1980). Most studies reported savannah sparrows nesting in undisturbed (idle) grasslands (Shields 1935, Graber and Graber 1963, Potter 1972, Owens and Myres 1973, Knight 1989, Sample 1989, Volkert 1992), meadows (Good and Dambach 1943, Brooks 1971, Dixon 1978, Wheelwright and Rising 1993), pastures (Graber and Graber 1963, Owens and Myres 1973, Sample 1989, Wheelwright and Rising 1993), hayfields (Graber and Graber 1963, Sample 1989, Wheelwright and Rising 1993), native and restored tallgrass prairie (Blankespoor 1980, Renken and Dinsmore 1987, Hartley 1994), short mixed-grass prairie (Stipa-Bouteloua) (Maher 1979), reclaimed surface mines (Whitmore and Hall 1978; Whitmore 1979a,b; Wray et al. 1982; Elliott 1991), CRP grasslands (Johnson and Schwartz 1993a,b; Reynolds et al. 1994; Swanson et al. 1995), roadside rights-of-way (Owens and Myres 1973, Warner 1992, Wheelwright and Rising 1993), grassed waterways in rowcrop fields (Bryan and Best 1991), and salt marsh-abandoned field ecotones (Bedard and LaPointe 1984, Zembal et al. 1988). Savannah sparrow nests also were found in small grains (Graber and Graber 1963, Sample 1989), no-tillage rowcrops (Basore et al. 1986), and cultivated rowcrops (Graber and Graber 1963, Sample 1989). Graber and Graber (1963) and Sample (1989) noted significantly higher densities of savannah sparrows in pastures and hayfields compared to small grains and rowcrops.
Wiens (1969) and Sample (1989) stated that savannah sparrows nested in areas of intermediate grass and forb cover, vegetation density, litter cover and depth, and herbaceous canopy height, and little woody cover. Wiens (1969) reported that 35 savannah sparrow nesting territories were characterized by 77% grass cover, 21% forb cover, 2% bare ground, and an average canopy height of 42 cm. Mean vegetative characteristics of habitats occupied by nesting savannah sparrows in Wisconsin were <1% woody cover, 75% herbaceous vegetation cover, 16% litter cover, 9% bare ground, 54-cm maximum vegetation height, and 17-cm vegetation height-density (Sample 1989). Savannah sparrow nesting territories on reclaimed surface mines in West Virginia were characterized by 7% grass cover, 11% forb cover, 58% litter cover, 24% bare ground, 1-cm litter depth, and 66-cm herbaceous canopy height (Whitmore 1979b).
Wiens (1969, 1973a) stated that savannah sparrows required dense, low vegetation for nesting. In Wisconsin, savannah sparrows avoided habitats with the tallest, densest vegetation and nested primarily in managed or disturbed habitats with shorter, sparser vegetation (Sample 1989). Sample (1989) reported a negative correlation between savannah sparrow abundance and maximum vegetation height and vegetation height-density. No difference in vegetation height was found between savannah sparrow nest sites and random points in Quebec, but successful nests were surrounded by taller vegetation than unsuccessful nests (Bedard and LaPointe 1984).
Nesting savannah sparrows showed a preference for grass-dominated habitats with little forb cover (Wiens 1969, 1973a; Welsh 1975; Knight 1989; Vickery et al. 1992). Sample (1989) and Swanson et al. (1995) reported a positive correlation between savannah sparrow abundance and percent herbaceous vegetation cover. In North Dakota, savannah sparrow abundance and percent grass cover were positively correlated (Renken and Dinsmore 1987). Savannah sparrows were most abundant on CRP fields with high percent grass and low percent legume cover in the north-central United States Johnson and Schwartz 1993b). However, in Oregon and Nevada, Rotenberry and Wiens (1980) found a positive correlation between savannah sparrow abundance and percent forb cover. In Maine, savannah sparrows nesting in areas of mainly forb and shrub cover experienced lower reproductive success than those nesting in predominately grass cover (Vickery et al. 1992).
Wiens (1969) noted that most savannah sparrow nesting territories were located in the center of grassland habitats, away from cultivated fields and fence lines. In West Virginia, however, Shields (1935) stated that savannah sparrows often included small trees, shrubs, and fence posts within the nesting territory. Although total woody cover in habitats used by savannah sparrows for nesting was low (<1%), Sample (1989) reported that the birds often used small trees and shrubs (<2 m tall), fence posts and wire, and tall herbaceous stems as singing perches. Wheelwright and Rising (1993) stated that savannah sparrows avoided nesting in areas with extensive woody cover.
Savannah sparrows foraged mainly outside the nesting territory in communal, nondefended areas (Dixon 1972, Welsh 1975). Foraging habitat was open and devoid of tall, dense vegetation (Wiens 1973a). During the breeding season adult savannah sparrows fed primarily on grasshoppers, beetles, and hemipterans (Welsh 1975, Rising 1987, Miller and McEwen 1995). Nestlings were fed almost exclusively insect larvae (Lepidoptera, Hymenoptera, and Diptera) and spiders (Welsh 1975, Meunier and Bedard 1984).
In Alberta, nesting densities of savannah sparrows were highest in undisturbed (> or equal to 3 growing seasons since last mowing) grasslands, but they also nested in grasslands that had been mowed the previous summer (Owens and Myres 1973). Herkert (1991) reported that savannah sparrow abundance was highest in recently mowed (within 4 months of 1 May) grasslands which provided the low-to-medium height vegetation preferred by this species. In Wisconsin, savannah sparrow nesting densities increased in hayfields after the first mowing (Sample 1989).
Several studies conducted in the Great Plains reported higher nesting densities of savannah sparrows on idle or lightly-grazed grasslands compared to moderately or heavily-grazed areas (Owens and Myres 1973, Maher 1979, Kantrud 1981, Kantrud and Kologiski 1982, Huber and Steuter 1984). For example, average breeding densities were 7.7 pairs/ha in ungrazed versus 0.3 pair/ha in grazed grasslands (Maher 1979). Moderately- and heavily-grazed areas did not provide the dense ground cover required by nesting savannah sparrows. Dale (1984) reported that although the number of nesting savannah sparrows was consistently lower in grazed than ungrazed areas, birds foraged most frequently in grazed grasslands. In Tennessee, Knight (1989) noted that savannah sparrows no longer used a grassy field after it was opened to heavy grazing. Herkert et al. (1993) recommended light grazing, resulting in > or equal to 40% vegetative cover > or equal to 25 cm tall, to produce the intermediate vegetation height and density preferred by savannah sparrows. Higher densities of brown-headed cowbirds were observed in summer-grazed versus winter-grazed pastures in Colorado, and Knopf et al. (1988) cautioned that summer grazing could lower reproductive success of savannah sparrows and other grassland-nesting birds.
In Minnesota (Tester and Marshall 1961), Wisconsin (Halvorsen and Anderson 1980), and South Dakota (Huber and Steuter 1984), higher nesting densities of savannah sparrows were found in unburned (for > or equal to 12 months) than burned grasslands. Halvorsen and Anderson (1980) attributed the difference in nesting densities to the lack of litter, required by nesting savannah sparrows, in burned areas. Tester and Marshall (1961) believed that savannah sparrows required > or equal to 2 years of litter accumulation post-fire before using a grassland for nesting. However, in Illinois, Herkert (1994a) reported that savannah sparrow densities were highest in grasslands the first growing season postfire, lower on grasslands the second growing season post-fire, and were not encountered on grasslands the third or more growing season post-fire. He concluded that savannah sparrows preferred recently burned grasslands for nesting.
Urbanization and the reversion of marginal farms to forest may have contributed to population declines of the savannah sparrow in the northeast and around the Great Lakes (Wheelwright and Rising 1993). Peterjohn (1989) and Peterjohn and Rice (1991) noted that savannah sparrows become scarce near large urban areas in Ohio. Sodhi (1992) compared savannah sparrow abundance between urban and rural areas in Canada and found them significantly more abundant in rural areas.
Based on data from Illinois and Missouri, Herkert et al. (1993) categorized the savannah sparrow as highly sensitive to habitat fragmentation. Species with a high area-sensitivity generally restrict nesting to large, contiguous habitats and rarely nest in small habitat fragments. In Illinois, Herkert (1991) found no savannah sparrows on grassland tracts <10 ha and noted that they were significantly more likely to occur on large than small grasslands. Savannah sparrow incidence increased with area and reached 50% at about 10 ha in Maine (Vickery et al. 1994) and 40 ha in Illinois (Herkert 1994c). Samson (1980) estimated that 1-10 ha represented the minimum amount of contiguous grassland habitat required to maintain a viable breeding population of savannah sparrows. In southwestern Ohio, 9 fields occupied by savannah sparrows averaged 15 ha (D. Nolin and J. Ritzenthaler, unpubl. data). Eight CRP grasslands occupied by savannah sparrows in Ohio averaged 9 ha and ranged from 3 to 13 ha (Oh. Div. Wildl., unpubl. data).