Skip to main content
NCBI home page
Ecology and Evolution logoLink to Ecology and Evolution
. 2024 Mar 13;14(3):e11099. doi: 10.1002/ece3.11099

Catastrophic storms, forest disturbance, and the natural history of Swainson's warbler

Gary R Graves 1,2,✉
PMCID: PMC10937492  PMID: 38487747

Abstract

The core breeding range of Swainson's warbler (Limnothlypis swainsonii) overlaps a zone of exceptionally high tornado frequency in southeastern North America. The importance of tornadoes in creating breeding habitat for this globally rare warbler and other disturbance‐dependent species has been largely overlooked. This paper estimates tornado frequency (1950–2021) and forest disturbance in the 240 counties and parishes in which breeding was documented from 1988 to 2014. The frequency of destructive tornadoes (EF1‐EF5) varied 6‐fold across the breeding range with a peak in the Gulf Coast states. Counties from east Texas to Alabama experienced the lowest median return interval of 5.4 years per 1000 km2, resulting in approximately 2477 ha of forest damage per 1000 km2 per century, based on current forestland cover. Tornadoes were significantly less frequent north and east of the core breeding range, with return intervals increasing to 9.1 years per 1000 km2 for breeding counties on the Atlantic coastal plain, 10.2 years per 1000 km2 in the Ozark Mountains, and 32.3 years per 1000 km2 in the Appalachian Mountains. Breeding counties within 150 km of the coastline from east Texas to North Carolina are also subjected to the highest frequency of hurricanes in the Western Hemisphere. Hurricanes often cause massive forest damage but archived meteorological and forestry data are insufficient to estimate the aggregate extent of forest disturbance in breeding counties. Nevertheless, the combined impact of tornadoes and hurricanes in the pre‐Anthropogenic era was likely sufficient to produce a dynamic mosaic of early‐successional forest crucial for the breeding ecology of Swainson's warbler. To ensure the long‐term survival of this rare warbler, it is advisable to develop habitat management plans that incorporate remote sensing data on early‐successional forest generated by catastrophic storms as well as anthropogenic activities.

Keywords: hurricanes, Limnothlypis swainsonii, Swainson's warbler, tornadoes

Swainson's warbler is a rare disturbance‐dependent songbird that breeds in forests of southeastern United States. This paper explores the role of catastrophic storms in creating breeding habitat in its contemporary breeding range.

graphic file with name ECE3-14-e11099-g002.jpg

1. INTRODUCTION

The serendipitous 1833 discovery of Swainson's warbler (Limnothlypis swainsonii) (Audubon, 1834) and Bachman's warbler (Vermivora bachmanii) (Audubon, 1833) by the Reverend John Bachman, near Parkers Ferry, South Carolina, sparked a question that has long perplexed ornithologists. Why are they so rare? These enigmatic southern warblers are linked by history, rarity, and geography. Their habitat preferences were apparently similar, with active nests of the two species discovered within 10 feet (3 m) of one another (Wayne, 1907). Bachman's warbler seemingly reached a population peak in the 1880s and 1890s when dozens were observed during spring migration in coastal Louisiana (Galbraith, 1888) and northern Florida (Brewster, 1891). However, this diminutive warbler was always regarded as a rare and locally‐distributed species and relatively little was discovered about its natural history while breeding individuals were still observable (Hamel, 1986; Stevenson, 1972). After a century of unexplained decline, it likely became extinct in the 1960s (Elphick et al., 2010), and was officially declared extinct in 2023 (U. S. Fish and Wildlife Service, 2023).

The extinction of Bachman's warbler was perhaps best explained by Terborgh (1989), who pointed out that extensive deforestation on its Cuban wintering grounds, coinciding with the 19th‐ and 20th‐ century‐expansion of the sugar industry, drastically reduced the carrying capacity during the nonbreeding season. The diminished population of males and females faced the insurmountable challenge of finding one another in the vastly larger breeding range that extended from the lower Mississippi Valley to coastal South Carolina. This problem was magnified by large‐scale logging of old‐growth bottomland forest following the Civil War through the 1930s (Williams, 1989). Mass cutting resulted in millions of hectares of early‐successional woodland and an overwhelming abundance of optimal breeding sites for yearlings, which typically settle many kilometers from their natal locations. Once the global population declined to a critical tipping point, natal dispersal quickly drove the species to extinction. The last few males observed in the 1950s and 1960s sang persistently but failed to attract mates (Terborgh, 1989).

Swainson's warbler could well face a similar fate in the coming century. This secretive species, which is nowhere common, now holds the unfortunate distinction of being the rarest breeding songbird on the mainland of southeastern North America (Figure 1). The sparsely distributed breeding population, recently estimated at 156,260 (Partners in Flight, 2022; Will et al., 2020), occupies a fragmented range that encompasses 1.14 million km2 from Texas to West Virginia (Partners in Flight Science Committee, 2013). The disappearance of peripheral populations in Delaware, Maryland, Missouri, and Illinois in recent decades signals a retraction toward the core breeding range in the Gulf Coast states. Population size is believed to be limited by deforestation on the breeding grounds (Graves, 2001, 2002) and wintering range in the Caribbean basin (Terborgh, 1989), in concert with changes in forest management policy (Graves & Tedford, 2016), flooding (Graves, 2001, 2002; Meanley, 1971; Reiley et al., 2013), and cowbird brood parasitism (Benson et al., 2010; Meanley, 1969).

FIGURE 1.

FIGURE 1

Open in a new tab

Swainson's warbler (Limnothlypis swainsonii). (a) Singing male (Red River Parish, Louisiana), (b) Lithograph of painting by John James Audubon (1834) of the type specimen collected by John Bachman in 1833. This specimen (USNM 2901) is deposited in the National Museum of Natural History, Smithsonian Institution, (c) Geolocated breeding territories in 240 counties and parishes from 1988 to 2014. Major rivers are highlighted in blue, county borders (parishes in Louisiana) in pale gray, and state boundaries in black. Inset map: AL (Alabama), AR (Arkansas), FL (Florida), GA (Georgia), KY (Kentucky), LA (Louisiana), MO (Missouri), MS (Mississippi), NC (North Carolina), OK (Oklahoma), SC (South Carolina), TN (Tennessee), TX (Texas), VA (Virginia), and WV (West Virginia).

Following its discovery, Swainson's warbler was virtually lost to science for 50 years until breeding sites were found in South Carolina and Georgia in the 1880s (Brewster, 1885). The resulting flurry of natural history notes focused on descriptions of nests and eggs, nest microhabitat, and geographic range extensions (Graves et al., 1996; Meanley, 1971), but contained little substantive information on landscape characteristics or the degree of natural and anthropogenic disturbance on breeding territories. The sole 19th century account from unequivocal old‐growth forest was documented in southeastern Missouri where Widmann (1895) found the warbler to be an uncommon breeding bird in bottomland forest where cypress and sweet gun towered over a mid‐story of hornbeam, ash and dogwood. The profusion of vine tangles and thickets of pawpaw, hazel, spicebush, and Hercules club at the site are indicative of disturbance and substantial light penetration to the understory. Regrettably, by the time the first quantitative habitat analysis was conducted by Eddleman et al. (1980), the opportunity to study the species in such primeval settings had largely passed. Consequently, our understanding of Swainson's warbler habitat selection is based mostly on studies in anthropogenically‐altered landscapes.

Contemporary populations of Swainson's warbler breed mostly in regenerating shelterwood cuts and clearcuts on commercial forestry lands where they reach their peak abundance. Smaller numbers also breed in a baffling variety of secondary habitat types (Bassett‐Touchell & Stouffer, 2006; Brooks & Legg, 1942; Graves, 2002, 2015, 2019, 2020; Henry, 2004; McNair, 2019; Meanley, 1971; Wood, 2014) that share a common characteristic—natural or anthropogenic disturbance, or a physiognomy that resembles the regenerative outcome of disturbance. Breeding territories invariably exhibit a high density of understory stems and/or foliage (Graves, 2002, 2015; Graves & Tedford, 2016) in patches sufficiently spacious to accommodate the warbler's expansive territories (Anich et al., 2009; Graves, 2001), which averaged 6.5 ha In the most intensively studied population in the lower Mississippi Valley. Habitat suitability is also influenced by flooding (Graves, 2001; Reiley, 2012), soil type, and soil moisture levels (Graves, 2002, 2015). These factors affect the deposition and decay of leaf litter and abundance and diversity of terrestrial litter arthropods (Brown, Benson, & Bednarz, 2011; Savage et al., 2010), which are essential resources for this dead‐leaf foraging specialist (Graves, 1998). The entire breeding distribution is characterized by moderate annual precipitation, and only a few peripheral breeding sites in Texas lie outside the 1000 mm‐isohyet.

A zone of exceptional tornado activity in the Gulf Coast states (Brooks et al., 2003; Gensini & Brooks, 2018), known as “Dixie Alley” among meteorologists (Dixon et al., 2011; Frazier et al., 2019; Gagan et al., 2010), extends from east Texas to Alabama. More than 75% of the global population of Swainson's warbler breeds in this geographic belt (Figure 2, Table 1). It is unclear whether this spatial overlap is coincidental or causative but field researchers are well aware of the warbler's affinity for disturbance gaps created by catastrophic storms. Despite this awareness, none of the 16 quantitative studies of breeding habitat has addressed the role of catastrophic storms in the formation of optimal breeding habitat (Bassett‐Touchell & Stouffer, 2006; Bednarz et al., 2005; Benson, 2008; Brown et al., 2009; Chartier, 2014; Eddleman et al., 1980; Graves, 2001, 2002; Graves & Tedford, 2016; Henry, 2004; Peters et al., 2005; Reiley et al., 2013; Somershoe et al., 2003; Thomas et al., 1996; Thompson, 2005; Wright, 2002). The omission stems in part from the limited size of study plots and the modest spatial scales of conservation programs. More significantly, it reflects the prevalence of anthropogenic disturbance and commercial forestry, which have supplanted natural events as the primary generators of early‐successional forest within the warbler's breeding range (Graves, 2015).

FIGURE 2.

FIGURE 2

Open in a new tab

Tornadoes in southeastern United States. (a) Tornado tracks (EF0‐EF5) in southeastern United States (1950–2021). Modified from https://www.spc.noaa.gov/gis/svrgis/images/tornado. (b) Tornado frequency (EF1–EF5; per 1000 km2 per century) in counties with documented breeding territories of Swainson's warbler.

TABLE 1.

Breeding population estimates for Swainson's warbler at the state level derived from Population Estimates Database, version 3.1 (Partners in Flight, 2022).

Alabama 45,370
Texas 31,100
Louisiana 16,100
Arkansas 13,000
Mississippi 13,000
North Carolina 9985
South Carolina 6700
Georgia 6340
Florida 5200
Tennessee 2775
Virginia 2490
West Virginia 2400
Kentucky 1800
Total 156,260
Open in a new tab

Note: Breeding populations in Oklahoma and Missouri were too small to warrant their inclusion.

This study characterizes the spatial gradients of tornado frequency in the southeastern United States and estimates tornado forest disturbance in context to the breeding ecology of Swainson's warbler. The analyses relied on three primary data sources. A geolocated archive of breeding territories (n = 1717) located in 240 counties and parishes across 15 states served as the geographic template for storm data analysis. Tornado data for these areas were obtained from NOAA's National Centers for Environmental Information (National Centers for Environmental Information, 2022a, 2022b). Additionally, county‐level data for forestland cover were acquired from the Forest Inventory and Analysis (FIA) database maintained by the USDA Forest Service (USDA Forest Service, 2022). These datasets facilitated the estimation of tornado frequency (Figure 2) and the extent of tornado forest disturbance in the contemporary breeding range (Figure 3).

FIGURE 3.

FIGURE 3

Open in a new tab

Tornado forest damage (ha per 1000 km2 per century). Estimates of county‐level damage were uncorrelated with area‐adjusted tornado frequency, because of extensive deforestation in some counties with high tornado frequencies (R s = 0.10, p = .11). Estimated forest damage exhibited a weak correlation with the percentage of forest cover despite the wide scatter of data points (R s = 0.16, p = .014).

A parallel analysis was conducted for hurricanes in southeastern United States (National Centers for Environmental Information, 2022a). Although the coastal‐to‐inland gradient in hurricane frequency is quite clear (Figure 4), county‐level estimates of hurricane forest damage could not be determined owing to data insufficiency. Both tornadoes and hurricanes are reviewed here despite their significant differences in scale and destructive potential. The resulting analyses provide valuable insight into the regional gradients of catastrophic storms and forest damage within the breeding range of Swainson's warbler. This information may be crucial for the effective management of this species and other disturbance‐dependent songbirds in the southeastern United States.

FIGURE 4.

FIGURE 4

Open in a new tab

Hurricane frequency (per 1000 km2 per century) in Swainson's warbler counties.

2. METHODS

2.1. Geolocation of territorial warblers

From 1988 to 2014, I surveyed breeding populations of Swainson's warbler in 15 states as part of a comprehensive study of the warbler's natural history (Graves, 1998, 2001, 2002, 2015, 2020; Graves & Tedford, 2016). These surveys targeted Swainson's warbler and were not incidental components of broader community censuses. Territorial males were documented in 240 counties and parishes by song recordings (see Data availability). I use the terms “warbler counties” and “breeding counties” interchangeably to refer to these administrative jurisdictions. Surveys were conducted during the breeding period, which began on 22 April in the Gulf Coast states and ended on 30 June in the Appalachian Mountains. I surveyed a wide spectrum of forestland and shrubland habitats, broadly classified as “forest land” by the USDA on public and private land and along waterways. Habitat physiognomy was sampled quantitatively on 7.2% of breeding territories, but these data were geographically clustered at only eight of hundreds of field sites (Graves, 2001, 2002; Graves & Tedford, 2016). Most breeding territories of this monogamous species were located using playback of songs, utilizing a protocol that was field‐tested and fine‐tuned in the late 1980s (Graves, 1996, 2001). Territorial males respond to playback by approaching the song source and delivering agitated “chip” notes, but usually refrain from singing until the playback source retreats or playback ends. Response to playback, mate‐guarding, persistence during “playback‐and‐follow” trials, and counter‐singing with other males were regarded as evidence of territoriality. Mist‐netting or other handling was not required to document territoriality. The geographic coordinates of territories were recorded on site with Garmin™ GPS receivers (post‐1998) or with Google Earth Pro from field notes and maps. A significant strength of this fieldwork archive is that it was compiled by a single senior investigator using comparable methods for 26 years. Data from other potential sources, such as eBird (https://ebird.org), were not used because of observer competency issues, geographic imprecision of checklists, the absence of photographic or auditory documentation, uncertainties about breeding status, and the widespread confusion between song variants of Swainson's warbler and Louisiana waterthrush (Parkesia motacilla).

2.2. Tornadoes: Overview and data

Tornado data spanning 71 years (1 January 1950 to 6 February 2021) for the 240 warbler counties were obtained from the NOAA database (National Centers for Environmental Information, 2022a). Tornado intensity was measured on the Fujita (F) scale until 2007 when the National Weather Service adopted the Enhanced Fujita (EF) scale (Doswell et al., 2009; Potter, 2007). The F and EF scales have six intensity categories indexed on damage to human structures and softwood and hardwood trees. Intensity scales are proxies for wind speed, which range from EFO (105–137 km h−1 for 3‐s gusts) to EF5 (>320 km h−1) (Edwards et al., 2013). Minimum wind speeds are slightly lower on the F scale than the EF scale for each category, e.g., F1 (126 km h−1) versus EF1 (137 km h−1). The frequency distribution of tornadoes by EF intensity scale appears to follow a power law in which the number of EF2 tornados is approximately half the number of EF1 tornadoes and so on (Elsner, Jagger, Widen, & Chavas, 2014). While low intensity tornadoes (F0 and EFO) are capable of stripping foliage and breaking small branches, they seldom result in windthrow or canopy damage that is visible by remote sensing (Zenoble & Peterson, 2017). EF1 (F1) tornadoes can uproot or topple mature trees, while EF2 (F2) and EF3 (F3) events can snap them off. High‐end EF3 (F3) tornadoes can debark trees, leaving only stubs of largest branches remaining on standing trunks. EF4 (F4) and EF5 (F5) tornadoes can cause catastrophic damage and are capable of reducing mature forests to barren debris fields.

The contiguous United States has experienced an average of 939 tornadoes per year (with a median of 915) from 1950 through 2021 (National Centers for Environmental Information, 2022b). The majority occurred east of the Rocky Mountains. Brooks et al. reported 495 tornadoes per year with an intensity rating of EF1 or higher from 1954 to 2013 (Brooks et al., 2014). In this study, I limited the analysis to intense tornadoes (>EF1 or F1) capable of creating large disturbance gaps in mature forests. An important limitation of the NOAA database is that it likely undercounts less severe tornadoes (EFO and EF1) in sparsely populated counties (Anderson et al., 2007; Potvin et al., 2022). Consequently, the analysis underestimates the actual frequency of tornadoes in southeastern United States. This study did not include data on derechos, downbursts, and microbursts, as county‐level records for these events were either unavailable or incomplete in the NOAA database. Although derechos are rarer than tornadoes, they can cause catastrophic forest disturbance equivalent to that generated by major tornadoes and hurricanes (Ashley & Mote, 2005; Bentley & Mote, 1998; Peterson, 2000).

Tornadoes in the NOAA database are tabulated by county. Average county area varies considerably among states: e.g., Alabama (x¯ = 2030 km2) versus Georgia (x¯ = 968 km2). Larger counties tend to experience more tornadoes than smaller counties, all else being equal (n = 240; R s = 0.58, p < .0001). The most intense tornadoes often cross county lines (Burow et al., 2020; Cannon et al., 2016), which amplifies spatial autocorrelation in county‐level counts of tornado occurrence. To mitigate the effects of county size, I standardized the tornado return frequency and return interval by county area (per 1000 km2 per century).

Smoothed, interpolated maps of tornado frequency in the continental United States, which include less intense tornados (EF0) over significantly shorter time intervals, are available in the meteorological literature (Brooks et al., 2003; Dixon et al., 2011; Frazier et al., 2019; Gensini & Brooks, 2018). However, using simple area‐standardized frequencies at the county level, as presented in this paper, offers transparency in both methodology and graphic output. This approach does not hinge on grid cell size, kernel density, or smoothing algorithms, providing a more straightforward and clear representation of tornado frequency in warbler counties.

2.3. Tornado track metrics

Tornado track length is defined as the distance traveled by tornadoes on the ground, while track width is the maximum‐recorded width of the tornado track. Median track length of tornadoes reported for the conterminous United States from 2007 to 2013 (Elsner, Jagger, & Elsner, 2014) increased monotonically with tornado intensity: EF1 (4.4 km, n = 2642), EF2 (10.0 km, n = 818), EF3 (23.3 km, n = 232), EF4 (34.8 km, n = 57), and EF5 (59.0 km, n = 9). The 1925 Tri‐State tornado, which started in Missouri and ended in Indiana (Johns et al., 2013), had a continuous track of 243 km (378 km including gaps). Median track width varies with tornado intensity: EF1 (91 m), EF2 (229 m), EF3 (549 m), EF4 (805 m), and EF5 (1920 m) (Elsner, Jagger, & Elsner, 2014). The 2013 El Reno tornado in Oklahoma holds the record for a maximum track width of 4.2 km (Bluestein et al., 2015). The ratio of median track length to width decreases with tornado intensity (EF1 = 48.0, EF2 = 43.9, EF3 = 42.4, EF4 = 43.3, and EF5 = 30.7).

I used several simplifying assumptions to estimate the areal footprint of forest disturbance. Assuming a rectangular track, the disturbance footprint can be roughly estimated by multiplying median track length by median track width (Cannon et al., 2016; Fujita, 1971). The width of individual tornado tracks often varies (Cannon et al., 2016; Karstens et al., 2013; Zenoble & Peterson, 2017), with patches of severe forest damage nested within less severely damaged patches that dissolve away from the center (Cannon et al., 2016). The degree of forest damage within the footprint will vary, from broken branches and stripped foliage at the track margins to toppled and snapped trees near the center. Tornado tracks often exhibit gaps (zero track width) with no severe forest damage. A study of 50 tornado tracks revealed a mean gap length of 13% of total track length (Zenoble & Peterson, 2017). I used this percentage as a correction factor in estimates of median footprint area for tornadoes: EF1 (35 ha), EF2 (199 ha), EF3 (1115 ha), EF4 (2439 ha), and EF5 events (9850 ha) (metadata from Elsner, Jagger, & Elsner, 2014). The mean frequency‐weighted footprint for all tornadoes (197 ha) was used to estimate forest disturbance gradients.

2.4. Forestland cover and tornado disturbance

Forest ecologists have conducted detailed studies on the impact of tornadoes on forest structure in eastern North America (Everham & Brokaw, 1996; Greenberg & Collins, 2016). These have focused on windthrow patterns produced by case events (Cannon et al., 2016; Peterson & Pickett, 1995), canopy structure and light penetration (Willson et al., 2020), and the resiliency of tree species to high winds (Peterson, 2007; Shirakura et al., 2006). My study aimed to provide a general approximation of historical rates of tornado forest damage in the 240 warbler counties by combining county‐level data for tornado records and forestland cover.

County‐level data for “forest land” coverage were obtained from the 2015 Forest Inventory and Analysis (FIA) database (USDA Forest Service, 2022). This database provides a coarse temporal snapshot of nominal forestland cover in the 240 warbler counties. Forestland cover varied widely, from 16 to 100% among the breeding counties, with nearly 6% of counties having >75% of land area classified as nonforest and more than 65% of counties having >25% of land area similarly classified. I used the median tornado footprint (197 ha), the area‐adjusted tornado frequency (tornadoes per 1000 km2 per century), and forestland coverage (%) to calculate rates of tornado forest damage (hectares per 1000 km2 per century) in the 240 warbler counties (Figure 3).

2.5. Hurricanes: Overview and data

County‐level data for hurricanes from 1 January 1996 to 1 April 2021 (25 years) were obtained from the NOAA National Centers for Environmental Information (National Centers for Environmental Information, 2022a). Hurricane intensity is scaled on the Saffir‐Simpson index ranging from Category 1 (wind speeds 119–153 km h−1) to Category 5 (wind speeds >252 km h−1) (https://www.nhc.noaa.gov/pdf/sshws.pdf). Category 1 hurricanes can snap large branches and topple shallowly rooted trees, while Category 2 storms can snap off mature trees. Major hurricanes (Category 3 or higher) may level mature forests. In general, economic damage to human infrastructure increases by a factor of four for every increase in category (Pielke et al., 2008) and forest disturbance may scale similarly with hurricane category. This study excluded tropical storms (<119 km h−1), although they can cause significant forest damage. Hurricanes at landfall are often >50 km in width (Irish et al., 2008), greater than the average widths of counties in the study area. However, because a hurricane may affect only a small portion of a county, I standardized hurricane incidence by county area (hurricanes per 1000 km2 per century; Figure 4), as was done for tornado data (Figure 2). County‐level estimates of hurricane forest damage could not be determined owing to the lack of county‐level data on the damage footprint of hurricane force winds.

2.6. Statistical analyses

County‐level tornado and hurricane frequencies were tested for goodness of fit to a normal distribution. The Lilliefors test showed significant deviation from normality for both tornado (D = 0.11, p = <.0001) and hurricane frequencies (D = 0.27, p < .0001), with right‐skewed and leptokurtic distributions. I therefore focused on median rather than mean values in statistical tests. Spearman rank correlation coefficients (r s) were used to evaluate the strength of correlations, while regional differences in tornado and hurricane frequency were evaluated with the Mann–Whitney U‐test. This test combines the distributions of two groups of values into a single sample and then assesses the range and location of the lowest group's distribution within the overall sample range against a ranked distribution that approaches normality (Hollander & Wolfe, 1973). Because one or both group sizes were larger than 20 in all comparisons, I used the normal approximation to calculate a z‐value for the Mann–Whitney U‐test. All p‐values are two‐tailed (α = 0.05). Statistical analyses were performed in R (version 4.2.0) with RStudio interface (version 2022.02.2) with the R Stats Package (RStudio Team, 2020).

3. RESULTS

3.1. Tornado return frequency

The 240 breeding counties examined in this study exhibited a median tornado frequency of 14 per 1000 km2 per century, with a range of 0 to 64 (Figure 2). Among the 145 breeding counties located within the meteorological “Dixie Alley,” which encompasses eastern Texas, southeastern Oklahoma, southern Arkansas, Louisiana, Mississippi, Alabama, and western Florida, and hosts the majority of breeding populations, the median tornado frequency was 18.5 tornadoes per 1000 km2 per century. Within this region, 127 of 145 counties (87.6%) reported a frequency of 10 or more tornadoes per 1000 km2 per century. The Ozark Mountains in Arkansas and Missouri, situated along the northern periphery of the breeding range, experienced a tornado frequency of 9.8 tornadoes per 1000 km2 per century, while counties in the Atlantic coastal plain reported a frequency of 11.0 tornadoes per 1000 km2 per century. The lowest tornado frequencies were observed in the Appalachian Mountains of West Virginia, Kentucky, Virginia, Tennessee, North Carolina, and South Carolina, with a median of 3.1 tornadoes per 1000 km2 per century.

3.2. Forest disturbance by tornadoes

Based on current percentages of forestland cover, tornado forest disturbance (Figure 3) in breeding counties was estimated to range from zero to 9656 ha per 1000 km2 per century (median = 2134 ha). These findings can also be interpreted as a forecast of future forest damage, assuming that there is no change in forestland area and if the frequency and intensity of tornadoes remain constant. Counties in the Gulf Coast states exhibited the highest rate of tornado forest damage (median = 2477 ha per 1000 km2 per century). This rate was significantly higher than estimates for the Atlantic coastal plain (median = 1461 ha; Mann–Whitney, Z = 6.27, p < .00001), Ozark Mountains (median = 1446 ha; Mann–Whitney, Z = 2.41, p = .016), and Appalachian Mountains (median = 470 ha; Mann–Whitney, Z = 7.28, p < .00001).

3.3. Hurricanes

The frequency of hurricanes in the 240 warbler counties ranged from zero to 40 per 1000 km2 per century, with a median of 1.6 per 1000 km2 per century. Hurricane force winds regularly extend 200 km inland along the Gulf coast, and occasionally to 400 km in the lower Mississippi Valley (Figure 4), but rarely more than 150 km inland on the Atlantic coast. The frequency of hurricanes in the Gulf Coast states (median = 3.9, n = 121 counties), including northwestern Florida, was similar to that observed on the Atlantic coastal plain (median = 3.3, n = 47; Mann–Whitney, Z = 0.25; p = .80). Overall, 28.8% (69 of 240) of counties experienced ≥5 hurricanes per 1000 km2 per century. The frequency of tornadoes and hurricanes was weakly correlated (r s = 0.15; p < .02; n = 240). Many inland areas with high tornado frequencies, such as eastern Arkansas and western Tennessee, experienced no hurricane force winds, while some Atlantic coastal counties with high hurricane frequencies experienced few tornadoes. The greatest concentration of severe storms, combining tornadoes and hurricanes, occurred in a geographic belt from Texas to Alabama and northwestern Florida, with a band of exceptionally high frequency that extended from the eastern parishes of Louisiana northward through Mississippi. The Appalachian region stood out as having a relatively low frequency of tornadoes and no hurricane‐force winds.

4. DISCUSSION

The core breeding range of Swainson's warbler spatially coincides with the hemispheric peaks of tornado frequency (Brooks et al., 2003; Dixon et al., 2011; Frazier et al., 2019; Gensini & Brooks, 2018) and hurricane incidence (Chambers et al., 2007; Doyle, 2009; Klotzbach et al., 2018; Pielke et al., 2008; Weinkle et al., 2012; Zeng et al., 2009). Estimating the correlation between storm frequency and the warbler's 19th century distribution and abundance is out of reach because of insufficient data. Evaluating the modern connection is equally challenging because there are relatively few sizeable forest tracts that are untouched by commercial forestry. However, the warbler's well‐documented response to forestry treatments provides compelling evidence of the crucial role that catastrophic storms once played in its breeding ecology. Commercial forestry is a functional analog of tornadoes and hurricanes in its capacity to thin or level extensive tracts of mature forest within short time frames. A comprehensive review of breeding habitat physiognomy suggests that, from the warbler's perspective, there may be little discernible difference between natural and anthropogenic forest disturbance at comparable scales.

The majority of contemporary Swainson's warbler populations are found on actively managed forestry lands (Eddleman et al., 1980; Graves, 2002; Peters et al., 2005; Twedt & Somershoe, 2009). The highest breeding concentrations were recorded in Louisiana (Graves, 2002) and South Carolina (Thompson, 2005), where densities reached 20.5 territories km−2 and 17 territories km−2, respectively. These figures narrowly exceed the theoretical maximum predicted from average territory size (6.5 ha) (Anich et al., 2009). Warblers colonize large disturbance gaps of anthropogenic or natural origin when regenerating vegetation attains the necessary structural complexity and understory density, typically >35,000 small woody stems ha−1 in deciduous woodland (Graves, 2002). The time lag between a forest disturbance event and warbler colonization depends on disturbance intensity, disturbance area, and vegetation growth rates. Large regenerating clearcuts in the Gulf Coast states typically provide suitable breeding habitat for 15–25 years, beginning 4–6 years after harvest (Graves, 2002; Peters et al., 2005; Thompson, 2005). Colonization trajectories in forest tracts leveled by strong tornadoes are likely comparable.

Breeding territories are characterized by understory thickets and vine tangles that provide visual screening for nesting and foraging sites (Graves, 2002; Graves & Tedford, 2016), as well as small semi‐concealed glades for terrestrial dead‐leaf foraging (Graves, 1998, 2002; Meanley, 1970). Warblers abandon territories when successional canopy closure shades out the understory (Graves & Tedford, 2016). Canopy height and floristics, per se, appear to exert little influence on habitat selection (Graves, 2002) as long as essential understory conditions are met. Soil characteristics, however, appear to influence both local and regional occupancy patterns (Graves, 2001, 2002, 2015). Nevertheless, the majority of outwardly suitable habitat tracts remain unoccupied, particularly on the periphery of the historic breeding range. This is a matter of considerable conservation concern.

Was it possible for tornadoes alone to generate enough forest disturbance to maintain breeding populations of Swainson's Warbler in pre‐anthropogenic landscapes? Acknowledging the pitfalls of extrapolating historical storm frequencies from recent NOAA data (National Centers for Environmental Information, 2022a), the median return interval for tornadoes (EF1–EF5) in the warbler counties in the core breeding range in the Gulf Coast states is approximately 5.4 year per 1000 km2. The return interval declines to 9.1 years per 1000 km2 on the Atlantic coastal plain, 10.2 years per 1000 km2 in the Ozark Mountains, and 32.3 years per 1000 km2 in the Appalachian Mountains. The observed 6‐fold decline in tornado frequency from the Gulf Coast states to the Appalachians represents the most significant meteorological gradient identified thus far within the circumscribed breeding distribution of the warbler.

Recent advances in remote sensing have made it possible to document tornado tracks in unprecedented detail, overcoming previous challenges in accurately measuring tornado forest damage (Burow et al., 2020; Cannon et al., 2016; Kingfield & de Beurs, 2017; Molthan et al., 2014; Rodríguez & Bech, 2020; Zenoble & Peterson, 2017). Case examples demonstrate the significant impact of exceptional tornadoes on forest structure. The most catastrophic tornado in recent history, in terms of forest destruction, occurred on 24 April 2010 (EF4) with a track length of 242 km that traversed seven counties in Mississippi (Wilkinson & Crosby, 2010). The tornado caused an estimated 16,700 ha of forest damage, with 3440 ha in Yazoo County alone (Wilkinson & Crosby, 2010). Although the probability of an EF1‐EF5 tornado strike at a randomly‐selected site in a given year is exceptionally low, the aggregate impact of hundreds of tornadoes each year within the warbler's core breeding range was likely sufficient to sustain an extensive patchwork of optimal breeding habitat prior to the extensive clearing of old‐growth forests during the 18th and 19th centuries.

The ecological implications of tornado track direction and geometry for disturbance‐dependent birds in forested landscapes have been overlooked by ecologists. The majority of tornadoes in southeastern North America propagate from the west and move in a northeastward direction (Suckling & Ashley, 2006), perpendicular to the south/north flyways of migratory songbirds (Figure 2a). Linear tornado tracks in forested areas may serve as ecological drift fences for disturbance‐dependent birds during migration and particularly during Lévy movements (Humphries et al., 2012; Reynolds, 2018; Viswanathan et al., 1996) in the pre‐mating and postnatal dispersal periods. Linear tornado tracks (with length‐to‐width ratios ranging from 30:1 to 50:1) likely facilitated the discovery of breeding territories and mates for low‐density species in pre‐anthropogenic landscapes, including the historically rare Swainson's warbler and Bachman's warbler. Unlike tornado tracks, other significant sources of forest disturbance such as hurricanes, fire, drought, ice storms, and insect outbreaks tend to leave footprints with more compact shapes.

The coastal plain from eastern Texas to North Carolina experiences the highest frequency of hurricanes in the Western Hemisphere (Cannon et al., 2023; Doyle, 2009; Weinkle et al., 2012). While hurricanes are less frequent than tornadoes, they cause significantly more damage per event owing to their greater size and longer duration (Peterson, 2000; Zeng et al., 2009). Forest impact models estimate that hurricanes destroyed an average of 147 million trees per year in the United States from 1851 to 1900, and 72 million trees per year from 1900 to 2000 (Zeng et al., 2009). The higher estimates before 1900 can be attributed to increased hurricane activity from 1870 to 1900, before many coastal regions were logged. The detrimental impact of hurricanes on avian species reliant on mature forest, notably the red‐cockaded woodpecker (Picoides borealis), has received considerable attention in this region (Lucash et al., 2022). In contrast, the collateral benefits for disturbance‐dependent species have been relatively understudied. Hurricane Katrina (see Figure 5), which made landfall in the Pearl River basin near the Louisiana‐Mississippi state line on 29 August 2005, presented a rare opportunity to directly document the response of Swainson's warbler to hurricane damage on a long‐term study plot (Brown et al., 2011). Following Katrina, its breeding density and that of several other species that prefer dense understory habitat increased significantly.

FIGURE 5.

FIGURE 5

Open in a new tab

Swainson's warbler territory located in hurricane‐damaged deciduous forest that was characterized by a closed canopy prior to Hurricane Katrina (29 August 2005) and Hurricane Gustav (1 September 2008). Photographed on 25 May 2009 in the Bogue Chitto National Wildlife Refuge, St. Tammany Parish, Louisiana.

In closing, a diverse assemblage of breeding birds in eastern North America is associated with disturbance gaps in predominately‐deciduous forests (Askins, 1998; Hunter et al., 2001), but no species appears to be more reliant on catastrophic storm damage than Swainson's warbler. Although most forestland in the warbler's contemporary breeding range in the Gulf Coast states is managed for woody biomass production on short rotation cycles, tornadoes and hurricanes continue to play a critical role in creating forest disturbance gaps of necessary size and frequency to support viable breeding populations of this and other disturbance‐dependent species. There is a pressing need to integrate remote sensing data on tornado and hurricane damage into regional habitat monitoring and management programs, particularly in public forestlands in the lower Mississippi Valley and Gulf coastal plain that are largely off‐limits to commercial forestry.

AUTHOR CONTRIBUTIONS

Gary R. Graves: Conceptualization (lead); data curation (lead); formal analysis (lead); funding acquisition (lead); investigation (lead); methodology (lead); project administration (lead); resources (lead); validation (lead); visualization (lead); writing – original draft (lead); writing – review and editing (lead).

CONFLICT OF INTEREST STATEMENT

The author declares no competing interests.

ACKNOWLEDGEMENTS

Robert Askins, T. J. Benson, William Eddleman, John Gerwin, Chuck Hunter, Chris Peterson, Scott Somershoe, and several anonymous reviewers provided insightful critiques of the manuscript. I acknowledge the continuing support of the Wetmore Fund (Smithsonian Institution) and the Smoketree Trust.

Graves, G. R. (2024). Catastrophic storms, forest disturbance, and the natural history of Swainson's warbler. Ecology and Evolution, 14, e11099. 10.1002/ece3.11099

DATA AVAILABILITY STATEMENT

All data generated or analyzed during this study are included in the Supporting Information file and the cited NOAA URLs. The Supporting Information file is deposited in Dryad (doi: 10.5061/dryad.gmsbcc2w8).

REFERENCES

  1. Anderson, C. J. , Wikle, C. K. , Zhou, Q. , & Royle, J. A. (2007). Population influences on tornado reports in the United States. Weather and Forecasting, 22, 571–579. [Google Scholar]
  2. Anich, N. M. , Benson, T. J. , & Bednarz, J. C. (2009). Estimating territory and home‐range sizes: Do singing locations alone provide an accurate estimate of space use? Auk, 126, 626–634. [Google Scholar]
  3. Ashley, W. S. , & Mote, T. L. (2005). Derecho hazards in the United States. Bulletin of the American Meteorological Society, 86(11), 1577–1592. 10.1175/bams-86-11-1577 [DOI] [Google Scholar]
  4. Askins, R. A. (1998). Restoring forest disturbances to sustain populations of shrubland birds. Restoration & Management Notes, 16(2), 166–173. [Google Scholar]
  5. Audubon, J. J. (1833). Birds of America (Folio) (p. 185). The Audubon Museum. [Google Scholar]
  6. Audubon, J. J. (1834). Ornithological biography (Vol. 2). Adam & Charles Black. [Google Scholar]
  7. Bassett‐Touchell, C. A. , & Stouffer, P. C. (2006). Habitat selection by Swainson's warblers breeding in loblolly pine plantations in southeastern Louisiana. Journal of Wildlife Management, 70, 1113–1119. [Google Scholar]
  8. Bednarz, J. C. , Stiller‐Krehel, P. , & Cannon, B. (2005). Distribution and habitat use of Swainson's warbler in eastern and northern Arkansas (pp. 576–588). USDA Forest Service General Technical Report PSW‐GTR‐191. [Google Scholar]
  9. Benson, T. J. (2008). Habitat use and demography of Swainson's warblers in eastern Arkansas. Ph.D. dissertation (p. 248). Arkansas State University. [Google Scholar]
  10. Benson, T. J. , Anich, N. M. , Brown, J. D. , & Bednarz, J. C. (2010). Habitat and landscape effects on brood parasitism, nest survival, and fledgling production in Swainson's warblers. Journal of Wildlife Management, 74, 81–93. [Google Scholar]
  11. Bentley, M. L. , & Mote, T. L. (1998). A climatology of derecho‐producing mesoscale convective systems in the central and eastern United States, 1986–95. Part I: Temporal and spatial distribution. Bulletin of the American Meteorological Society, 79(11), 2527–2540. [Google Scholar]
  12. Bluestein, H. B. , Snyder, J. C. , & Houser, J. B. (2015). A multiscale overview of the El Reno, Oklahoma, tornadic supercell of 31 may 2013. Weather and Forecasting, 30, 525–552. [Google Scholar]
  13. Brewster, W. (1885). Swainson's warbler. Auk, 2, 65–80. [Google Scholar]
  14. Brewster, W. (1891). Notes on Bachman's warbler (Helminthophila bachmani). Auk, 8, 149–157. [Google Scholar]
  15. Brooks, H. E. , Carbin, G. W. , & Marsh, P. T. (2014). Increased variability of tornado occurrence in the United States. Science, 346(6207), 349–352. 10.1126/science.1257460 [DOI] [PubMed] [Google Scholar]
  16. Brooks, H. E. , Doswell, C. A. , & Kay, M. P. (2003). Climatological estimates of local daily tornado probability for the United States. Weather and Forecasting, 18(4), 626–640. [DOI] [Google Scholar]
  17. Brooks, M. , & Legg, W. C. (1942). Swainson's warbler in Nicholas County, West Virginia. Auk, 59, 76–86. [Google Scholar]
  18. Brown, D. R. , Sherry, T. W. , & Harris, J. (2011). Hurricane Katrina impacts the breeding bird community in a bottomland hardwood forest of the Pearl River basin, Louisiana. Forest Ecology and Management, 261(1), 111–119. 10.1016/j.foreco.2010.09.038 [DOI] [Google Scholar]
  19. Brown, J. D. , Benson, T. J. , & Bednarz, J. C. (2009). Vegetation characteristics of Swainson's warbler habitat at the White River National Wildlife Refuge, Arkansas. Wetlands, 29, 586–597. [Google Scholar]
  20. Brown, J. D. , Benson, T. J. , & Bednarz, J. C. (2011). Arthropod communities associated with habitats occupied by breeding Swainson's warblers. Condor, 113, 890–898. 10.1525/cond.2011.100184 [DOI] [Google Scholar]
  21. Burow, D. , Herrero, H. V. , & Ellis, K. N. (2020). Damage analysis of three long‐track tornadoes using high‐resolution satellite imagery. Atmosphere, 11(6), 613. [Google Scholar]
  22. Cannon, J. B. , Hepinstall‐Cymerman, J. , Godfrey, C. M. , & Peterson, C. J. (2016). Landscape‐scale characteristics of forest tornado damage in mountainous terrain. Landscape Ecology, 31(9), 2097–2114. 10.1007/s10980-016-0384-8 [DOI] [Google Scholar]
  23. Cannon, J. B. , Peterson, C. J. , Godfrey, C. M. , & Whelan, A. W. (2023). Hurricane wind regimes for forests of North America. Proceedings of the National Academy of Sciences, 120(42), e2309076120. 10.1073/pnas.2309076120 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Chambers, J. Q. , Fisher, J. I. , Zeng, H. , Chapman, E. L. , Baker, D. B. , & Hurtt, G. C. (2007). Hurricane Katrina's carbon footprint on U.S. gulf coast forests. Science, 318(5853), 1107. 10.1126/science.1148913 [DOI] [PubMed] [Google Scholar]
  25. Chartier, N. A. (2014). Breeding biology of Swainson's warbler (Limnothlypis swainsonii) in a North Carolina bottomland hardwood forest. Ph.D. dissertation (p. 147). North Carolina State University. [Google Scholar]
  26. Dixon, P. G. , Mercer, A. E. , Choi, J. , & Allen, J. S. (2011). Tornado risk analysis: Is Dixie alley an extension of tornado alley? Bulletin of the American Meteorological Society, 92(4), 433–441. 10.1175/2010bams3102.1 [DOI] [Google Scholar]
  27. Doswell, C. A. , Brooks, H. E. , & Dotzek, N. (2009). On the implementation of the enhanced Fujita scale in the USA. Atmospheric Research, 93(1), 554–563. 10.1016/j.atmosres.2008.11.003 [DOI] [Google Scholar]
  28. Doyle, T. W. (2009). Hurricane frequency and landfall distribution for coastal wetlands of the Gulf coast, USA. Wetlands, 29(1), 35–43. 10.1672/08-36.1 [DOI] [Google Scholar]
  29. Eddleman, W. R. , Evans, K. E. , & Elder, W. H. (1980). Habitat characteristics and management of Swainson's warbler in southern Illinois. Wildlife Society Bulletin, 8, 228–233. [Google Scholar]
  30. Edwards, R. , LaDue, J. G. , Ferree, J. T. , Scharfenberg, K. , Maier, C. , & Coulbourne, W. L. (2013). Tornado intensity estimation: Past, present, and future. Bulletin of the American Meteorological Society, 94, 641–653. [Google Scholar]
  31. Elphick, C. S. , Roberts, D. L. , & Reed, J. M. (2010). Estimated dates of recent extinctions for north American and Hawaiian birds. Biological Conservation, 143, 617–624. [Google Scholar]
  32. Elsner, J. B. , Jagger, T. H. , & Elsner, I. J. (2014). Tornado intensity estimated from damage path dimensions. PLoS One, 9(9), e107571. 10.1371/journal.pone.0107571 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Elsner, J. B. , Jagger, T. H. , Widen, H. M. , & Chavas, D. R. (2014). Daily tornado frequency distributions in the United States. Environmental Research Letters, 9(2), 024018. 10.1088/1748-9326/9/2/024018 [DOI] [Google Scholar]
  34. Everham, E. M. , & Brokaw, N. V. L. (1996). Forest damage and recovery from catastrophic wind. The Botanical Review, 62(2), 113–185. 10.1007/BF02857920 [DOI] [Google Scholar]
  35. Frazier, A. E. , Hemingway, B. L. , & Brasher, J. P. (2019). Land surface heterogeneity and tornado occurrence: An analysis of tornado alley and Dixie alley. Geomatics, Natural Hazards and Risk, 10(1), 1475–1492. 10.1080/19475705.2019.1583292 [DOI] [Google Scholar]
  36. Fujita, T. T. (1971). Proposed characterization of tornadoes and hurricanes by area and intensity. Satellite and Mesometeorology Research Project Research Paper No. 91. [Google Scholar]
  37. Gagan, J. P. , Gerard, A. , & Gordon, J. (2010). A historical and statistical comparison of “tornado alley” to “Dixie alley”. National Weather Digest, 34, 145–155. [Google Scholar]
  38. Galbraith, C. S. (1888). Bachman's warbler (Helminthophila bachmani) in Louisiana. Auk, 5, 323. [Google Scholar]
  39. Gensini, V. A. , & Brooks, H. E. (2018). Spatial trends in United States tornado frequency. npj Climate and Atmospheric Science, 1(1), 38. 10.1038/s41612-018-0048-2 [DOI] [Google Scholar]
  40. Graves, G. R. (1996). Censusing wintering populations of Swainson's warbler: Surveys in the Blue Mountains of Jamaica. Wilson Bulletin, 108, 94–103. [Google Scholar]
  41. Graves, G. R. (1998). Stereotyped foraging behavior of the Swainson's warbler. Journal of Field Ornithology, 69, 121–127. [Google Scholar]
  42. Graves, G. R. (2001). Factors governing the distribution of Swainson's warbler along a hydrological gradient in great dismal swamp. Auk, 118, 650–664. [Google Scholar]
  43. Graves, G. R. (2002). Habitat characteristics in the core breeding range of the Swainson's warbler. Wilson Bulletin, 114, 210–220. [Google Scholar]
  44. Graves, G. R. (2015). Recent large‐scale colonisation of southern pine plantations by Swainson's warbler Limnothlypis swainsonii . Bird Conservation International, 2015, 1–14. [Google Scholar]
  45. Graves, G. R. (2019). Swainson's warbler (Limnothlypis swainsonii) in a monoculture of invasive Japanese knotweed (Reynoutria japonica). Northeastern Naturalist, 26, N9–N11. [Google Scholar]
  46. Graves, G. R. (2020). Incidence of invasive privet (Ligustrum spp.) on Swainson's warbler breeding territories in the Gulf coastal states. Southeastern Naturalist, 19(4), 627–636. [Google Scholar]
  47. Graves, G. R. , Simpson, M. B. , & Stephens, L. D. (1996). Joseph LeConte and the discovery of the nest of Swainson's warbler (Limnothlypis swainsonii): A reconsideration. Archives of Natural History, 23, 287–290. [Google Scholar]
  48. Graves, G. R. , & Tedford, B. L. (2016). Common denominators of Swainson's warbler habitat in bottomland hardwood forest in the White River watershed in southeastern Arkansas. Southeastern Naturalist, 15, 315–330. [Google Scholar]
  49. Greenberg, C. H. , & Collins, B. S. (Eds.). (2016). Natural disturbances and historic range of variation: Type, frequency, severity, and post‐disturbance structure in central hardwood forests USA. Springer. [Google Scholar]
  50. Hamel, P. B. (1986). Bachman's warbler, a species in peril. Smithsonian Institution Press. [Google Scholar]
  51. Henry, D. R. (2004). Reproductive success and habitat selection of Swainson's warblers in managed pine versus bottomland hardwood forests. Ph.D. dissertation (p. 139). Tulane University. [Google Scholar]
  52. Hollander, M. , & Wolfe, D. A. (1973). Nonparametric statistical methods. John Wiley and Sons. [Google Scholar]
  53. Humphries, N. E. , Weimerskirch, H. , Queiroz, N. , Southall, E. J. , & Sims, D. W. (2012). Foraging success of biological Lévy flights recorded in situ. Proceedings of the National Academy of Sciences, 109(19), 7169–7174. 10.1073/pnas.1121201109 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Hunter, W. C. , Buehler, D. A. , Canterbury, R. A. , Confer, J. L. , & Hamel, P. B. (2001). Conservation of disturbance‐dependent birds in eastern North America. Wildlife Society Bulletin, 29, 440–455. [Google Scholar]
  55. Irish, J. L. , Resio, D. T. , & Ratcliff, J. J. (2008). The influence of storm size on hurricane surge. Journal of Physical Oceanography, 38(9), 2003–2013. 10.1175/2008jpo3727.1 [DOI] [Google Scholar]
  56. Johns, R. H. , Burgess, D. W. , Gilmore, M. S. , Hart, J. A. , & Piltz, S. F. (2013). The 1925 tri‐state tornado damage path and associated storm system. Electronic Journal of Severe Storms Meteorology, 8, 1–33. [Google Scholar]
  57. Karstens, C. D. , Gallus, W. A., Jr. , Lee, B. D. , & Finley, C. A. (2013). Analysis of tornado‐induced tree fall using aerial photography from the Joplin, Missouri, and Tuscaloosa–Birmingham, Alabama, tornadoes of 2011. Journal of Applied Meteorology and Climatology, 52, 1049–1068. [Google Scholar]
  58. Kingfield, D. M. , & de Beurs, K. M. (2017). Landsat identification of tornado damage by land cover and an evaluation of damage recovery in forests. Journal of Applied Meteorology and Climatology, 56(4), 965–987. 10.1175/jamc-d-16-0228.1 [DOI] [Google Scholar]
  59. Klotzbach, P. J. , Bowen, S. G. , Pielke, R. , & Bell, M. (2018). Continental U.S. hurricane landfall frequency and associated damage: Observations and future risks. Bulletin of the American Meteorological Society, 99(7), 1359–1376. 10.1175/bams-d-17-0184.1 [DOI] [Google Scholar]
  60. Lucash, M. S. , Weiss, S. , Duveneck, M. J. , & Scheller, R. M. (2022). Managing for red‐cockaded woodpeckers is more complicated under climate change. Journal of Wildlife Management, 86(8), e22309. 10.1002/jwmg.22309 [DOI] [Google Scholar]
  61. McNair, D. B. (2019). Swainson's warbler breeding distribution and habit characteristics in bottomland hardwood forests of the lower piedmont in North Carolina: Importance of Chinese privet. Southeastern Naturalist, 18, 510–524. [Google Scholar]
  62. Meanley, B. (1969). Pre‐nesting and nesting behavior of the Swainson's warbler. Wilson Bulletin, 81, 246–257. [Google Scholar]
  63. Meanley, B. (1970). Method of searching for food by the Swainson's warbler. Wilson Bulletin, 82, 228. [Google Scholar]
  64. Meanley, B. (1971). Natural history of the Swainson's warbler. North America Fauna, 69, 1–90. [Google Scholar]
  65. Molthan, A. L. , Bell, J. R. , Cole, T. A. , & Burks, J. E. (2014). Satellite‐based identification of tornado damage tracks from the 27 April 2011 severe weather outbreak. Journal of Operational Meteorology, 2, 191–208. [Google Scholar]
  66. National Centers for Environmental Information . (2022a). Storm events database. National Oceanic and Atmospheric Association (NOAA). [Google Scholar]
  67. National Centers for Environmental Information . (2022b). U.S. tornadoes. National Oceanic and Atmospheric Association (NOAA). [Google Scholar]
  68. Partners in Flight . (2022). Population estimates database, version 3.1. https://pif.birdconservancy.org/population‐estimate‐database‐scores/
  69. Partners in Flight Science Committee . (2013). Population estimates database, version 2013. http://rmbo.org/pifpopestimates/Database.aspx
  70. Peters, K. A. , Lancia, R. A. , & Gerwin, J. A. (2005). Swainson's warbler habitat selection in a managed bottomland hardwood forest. Journal of Wildlife Management, 69, 409–417. [Google Scholar]
  71. Peterson, C. J. (2000). Catastrophic wind damage to north American forests and the potential impact of climate change. Science of the Total Environment, 262(3), 287–311. 10.1016/S0048-9697(00)00529-5 [DOI] [PubMed] [Google Scholar]
  72. Peterson, C. J. (2007). Consistent influence of tree diameter and species on damage in nine eastern North America tornado blowdowns. Forest Ecology and Management, 250(1), 96–108. 10.1016/j.foreco.2007.03.013 [DOI] [Google Scholar]
  73. Peterson, C. J. , & Pickett, S. T. A. (1995). Forest reorganization: A case study in an old‐growth forest catastrophic blowdown. Ecology, 76(3), 763–774. 10.2307/1939342 [DOI] [Google Scholar]
  74. Pielke, R. A. , Gratz, J. , Landsea, C. W. , Collins, D. , Saunders, M. A. , & Musulin, R. (2008). Normalized hurricane damage in the United States: 1900–2005. Natural Hazards Review, 9, 29–42. [Google Scholar]
  75. Potter, S. (2007). Fine‐tuning Fujita: After 35 years, a new scale for rating tornadoes takes effect. Weatherwise, 60(2), 64–71. 10.3200/WEWI.60.2.64-71 [DOI] [Google Scholar]
  76. Potvin, C. K. , Broyles, C. , Skinner, P. S. , & Brooks, H. E. (2022). Improving estimates of U.S. tornado frequency by accounting for unreported and underrated tornadoes. Journal of Applied Meteorology and Climatology, 61(7), 909–930. 10.1175/jamc-d-21-0225.1 [DOI] [Google Scholar]
  77. Reiley, B. M. (2012). The effects of a riverine flood on the ecology and habitat of the Swainson's warbler and a test of the Swainson's warbler habitat suitability index. M.Sc. thesis (p. 113). Arkansas State University. [Google Scholar]
  78. Reiley, B. M. , Benson, T. J. , & Bednarz, J. C. (2013). Mechanisms of flood‐inducing territory abandonment in an obligate ground‐foraging bird. Condor, 115, 650–658. [Google Scholar]
  79. Reynolds, A. M. (2018). Current status and future directions of Lévy walk research. Biology Open, 7(1), 30106. 10.1242/bio.030106 [DOI] [PMC free article] [PubMed] [Google Scholar]
  80. Rodríguez, O. , & Bech, J. (2020). Reanalysing strong‐convective wind damage paths using high‐resolution aerial images. Natural Hazards, 104(1), 1021–1038. 10.1007/s11069-020-04202-6 [DOI] [Google Scholar]
  81. RStudio Team . (2020). RStudio: Integrated development for R. RStudio. [Google Scholar]
  82. Savage, A. L. , Moorman, C. E. , Gerwin, J. A. , & Sorenson, C. (2010). Prey selection by Swainson's warblers on the breeding grounds. Condor, 112, 605–614. 10.1525/cond.2010.090055 [DOI] [Google Scholar]
  83. Shirakura, F. , Sasaki, K. , Arévalo, J. R. , & Palmer, M. W. (2006). Tornado damage of Quercus stellata and Quercus marilandica in the cross timbers, Oklahoma, USA. Journal of Vegetation Science, 17(3), 347–352. 10.1111/j.1654-1103.2006.tb02454.x [DOI] [Google Scholar]
  84. Somershoe, S. G. , Hudman, S. P. , & Chandler, C. R. (2003). Habitat use by Swainson's warblers in a managed bottomland forest. Wilson Bulletin, 115, 148–154. [Google Scholar]
  85. Stevenson, H. M. (1972). The recent history of Bachman's warbler. Wilson Bulletin, 84(3), 344–347. [Google Scholar]
  86. Suckling, P. W. , & Ashley, W. S. (2006). Spatial and temporal characteristics of tornado path direction. Professional Geographer, 58(1), 20–38. 10.1111/j.1467-9272.2006.00509.x [DOI] [Google Scholar]
  87. Terborgh, J. (1989). Where have all the birds gone? Princeton University Press. [Google Scholar]
  88. Thomas, B. G. , Wiggers, E. P. , & Clawson, R. L. (1996). Habitat selection and breeding status of Swainson's warbler in southern Missouri. Journal of Wildlife Management, 60, 611–616. [Google Scholar]
  89. Thompson, J. L. (2005). Breeding biology of Swainson's warblers in a managed South Carolina bottomland forest. Ph.D. dissertation (p. 164). North Carolina State University. [Google Scholar]
  90. Twedt, D. J. , & Somershoe, S. G. (2009). Bird response to prescribed silvicultural treatments in bottomland hardwood forests. Journal of Wildlife Management, 73, 1140–1150. [Google Scholar]
  91. U. S. Fish and Wildlife Service . (2023). Endangered and threatened wildlife and plants; removal of 21 extinct species from the list of endangered and threatened wildlife. Federal Register, 88, 71644–71682. [Google Scholar]
  92. USDA Forest Service . (2022). Forest biomass across the lower 48 states and Alaska. U.S. Department of Agriculture (USDA) Forest Service. [Google Scholar]
  93. Viswanathan, G. M. , Afanasyev, V. , Buldyrev, S. V. , Murphy, E. J. , Prince, P. A. , & Stanley, H. E. (1996). Lévy flight search patterns of wandering albatrosses. Nature, 381(6581), 413–415. 10.1038/381413a0 [DOI] [PubMed] [Google Scholar]
  94. Wayne, A. T. (1907). The nest and eggs of Bachman's warbler, Helminthophila bachmani (Aud.), taken near Charleston, South Carolina. Auk, 44, 43–48. [Google Scholar]
  95. Weinkle, J. , Maue, R. , & Pielke, R. (2012). Historical global tropical cyclone landfalls. Journal of Climate, 25(13), 4729–4735. 10.1175/jcli-d-11-00719.1 [DOI] [Google Scholar]
  96. Widmann, O. (1895). Swainson's warbler an inhabitant of the swampy woods of southeastern Missouri. Auk, 12, 112–117. [Google Scholar]
  97. Wilkinson, D. W. , & Crosby, M. K. (2010). Rapid assessment of forest damage from tornadoes in Mississippi. Photogrammetric Engineering and Remote Sensing, 76, 1298–1301. [Google Scholar]
  98. Will, T. , Stanton, J. C. , Rosenberg, K. V. , Panjabi, A. O. , Camfield, A. F. , & Shaw, A. E. (2020). Handbook to the Partners in Flight Population Estimates Database, Version 3.1. PIF Technical Series No 7.1 . pif.birdconservancy.org/popest.handbook.pd
  99. Williams, M. (1989). Americans and their forests: A historical geography. Cambridge University Press. [Google Scholar]
  100. Willson, K. G. , Cox, L. E. , Hart, J. L. , & Dey, D. C. (2020). Three‐dimensional light structure of an upland Quercus stand post‐tornado disturbance. Journal of Forestry Research, 31(1), 141–153. 10.1007/s11676-019-00907-y [DOI] [Google Scholar]
  101. Wood, D. M. (2014). Swainson's warbler and breeding bird atlas surveys in south‐central West Virginia 2013–2014. Cooperative project #13‐R06 final report. West Virginia Department of Natural Resources. [Google Scholar]
  102. Wright, E. A. (2002). Breeding population density and habitat use of Swainson's warbler in a Georgia floodplain forest. M.Sc. thesis (p. 89). University of Georgia. [Google Scholar]
  103. Zeng, H. , Chambers, J. Q. , Negrón‐Juárez, R. I. , Hurtt, G. C. , Baker, D. B. , & Powell, M. D. (2009). Impacts of tropical cyclones on U.S. forest tree mortality and carbon flux from 1851 to 2000. Proceedings of the National Academy of Sciences, 106(19), 7888–7892. 10.1073/pnas.0808914106 [DOI] [PMC free article] [PubMed] [Google Scholar]
  104. Zenoble, M. D. , & Peterson, C. J. (2017). Remotely visible width and discontinuity of 50 tornado damage paths through forested landscapes. Electronic Journal of Severe Storms Meteorology, 12, 1–21. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

All data generated or analyzed during this study are included in the Supporting Information file and the cited NOAA URLs. The Supporting Information file is deposited in Dryad (doi: 10.5061/dryad.gmsbcc2w8).

Articles from Ecology and Evolution are provided here courtesy of Wiley

RESOURCES