Grassland and Shrubland Birds within the Carden Plain: Recent Monitoring Results

Transcription

1 Grassland and Shrubland Birds within the Carden Plain: Recent Monitoring Results 1

2 Grassland and Shrubland Birds within the Carden Plain: Recent Monitoring Results Compiled by Ron Reid January 2011 Summary of Conclusions: 1. Birds of the Carden Plain are not exempt from the documented regional declines in grassland species, with average annual declines of 2% to 4% over the past decade, even though habitat conditions and extent do not appear to have changed significantly. 2. Open grassland habitats support the highest density of target bird species, and have maintained the highest frequency of area-sensitive species. 3. Changes in abundance of individual species show a pronounced trend towards losses of short-grass specialists such as Horned Lark, and increases in a small group of shrubland species, notably Song Sparrow and Eastern Towhee. 4. Diversity of shrubland species is showing a strong decline, at the same time as the abundance of two shrubland species is increasing rapidly. 5. Annual monitoring shows declines for 9 grassland species and 5 of 7 shrubland species, but with considerable unexplained variation between Cameron Ranch and Windmill Ranch plots. 6. Habitat patches in Ramara Township consistently rank lower in target bird diversity and density than the Carden IBA plots to the east, with that pattern especially pronounced for Upland Sandpiper and Eastern Meadowlark, both area-sensitive species. 7. The most significant grassland areas identified in Ramara are clustered southeast of Brechin and north of the Trent-Severn Canal. 8. The extensive grasslands of the CKL South section (north half of former Eldon Township) stand out as particularly diverse and productive in their bird life. 2

3 Acknowledgements Major parts of this manuscript have been summarized from data analysis and internal reports prepared by Kristen Field and Kyra Howes of The Couchiching Conservancy. Financial support has been provided by the Ontario Species at Risk Stewardship Fund, Environment Canada Habitat Stewardship Program, and the George Cedric Metcalf Charitable Foundation. Data collection for these projects would not have been possible without the dedication of dozens of volunteers from Ontario Field Ornithologists, Toronto Ornithological Club, Carden Field Naturalists, and Couchiching Conservancy. Their assistance is greatly appreciated. The Couchiching Conservancy is a non-profit, charitable land trust whose mission is to protect the special natural features of the Couchiching region for future generations. The Ontario Species at Risk Stewardship Fund supports greater public involvement in all kinds of species at risk protection and recovery activities. The fund provides $18 million to eligible projects over four years ( ). The purpose of the fund is to inspire and encourage people to become involved in species at risk recovery, and build on the good work already done by groups and individuals. The Government of Canada Habitat Stewardship Program for Species at Risk. Funds from this program promote the participation of local communities to help with the recovery of species at risk and prevent other species from becoming a conservation concern. 3

4 Background The Carden Plain lies northeast of Lake Simcoe on an area of shallow limestone bedrock (Chapman and Putnam, 1984). Some sections of this landform have developed alvar habitats characterized by extremely thin soils and distinctive ecological communities. Other extensive areas were cleared in the late 1800s for cattle ranching, and have been maintained in this use as rough pasture set between linear wooded wetlands. The shallow droughty soils in this area have prevented conversion to agricultural cropland, with the exception of a few scattered hayfields. These conditions, together with frequent wildfires until the late 1940s, have also limited the growth of large woody species, although the amount of low shrubland has increased in recent decades. Therefore, the Carden Plain retains a significant amount of natural grassland habitat, and is a haven for relatively large populations of grassland birds. The Carden limestone plain has been recognized for several decades as an exceptional breeding habitat for grassland and shrubland birds. In 1998, Birdlife International recognized most of former Carden Township and parts of the adjacent townships of Eldon, Bexley, and Ramara as an Important Bird Area (IBA) of national significance (Coxon and Reid 2001). In 2008, the Couchiching Conservancy and partner groups published an Integrated Carden Conservation Strategy which addressed the balance among conservation, quarry, and community values in this area (Couchiching Conservancy, 2008). A more detailed analysis of conservation values, published in 2009 as the Carden Conservation Plan Working Draft, recognized that the Carden limestone plain extends westward well into Ramara Township (Eco Source Management 2009). The Plan identified three sections within the Carden Plain North Kawartha Lakes portion (from the Trent-Severn Waterway northwards), South Kawartha Lakes portion (primarily in former Eldon township), and the Ramara portion (see Figure 1). The Conservation Plan noted that there is relatively little background information available for the Ramara section, and further background and baseline studies need to be conducted there. Grassland birds are rapidly declining throughout North America over the past several decades (Cadman et. al. 2007; Sauer et al. 2005). Indeed, Bird Studies Canada has concluded that grassland birds in North America have experienced the most pronounced declines of any group of birds on the continent, and the declines appear to be continuing unabated. (McCracken, 2005; BSC, 2009). Among 37 species of grassland birds monitored by the Breeding Bird Survey, 32 are showing some form of decline, while only 5 are increasing (Sauer et al. 2005). The most pronounced declines are taking place in the north-eastern sections of the range of grassland birds on the continent (Jon McCrackin, pers. comm.) 4

5 Table 1: High-priority grassland bird species in southern Ontario, with Breeding Bird Survey population trends (Bird Studies Canada 2009) 1. E = endangered SC = Special Concern Southern Ontario is an important area for grassland birds because of their relatively high representation in this area. In fact, it has been recognized as the largest and most important area of grassland in the Northeast by the Partners in Flight program (Pashley et. al., 2000). For example, Bird Studies Canada notes that Ontario supports one-fifth of the World s population of Bobolinks (Heagy and McCracken, 2004). Considerable conservation effort has been undertaken on the Carden Plain since 2001, both to acquire significant habitats and to encourage good stewardship of landscapes of conservation interest. During the same period, additional properties have been acquired by the aggregate industry, with several major new sites licensed as future quarries. However, to date there has been very little obvious change in actual land uses or habitat conditions, with the exception of a few small sites that have been excluded from this study. 5

6 6

7 Sources of Monitoring Data A. IBA monitoring comparison In the year 2000, a monitoring program based on standardized point counts was carried out to identify high priority grassland and shrubland habitats within the Carden Plain IBA. Each relatively homogenous area of habitat was delineated and given a Habitat Patch number. Permanent monitoring plots for each Habitat Patch were set by recording the UTM coordinates; most of the sites fell along accessible roadsides. Each Patch was assigned to a habitat type and the size of each Habitat Patch was estimated. A second round of monitoring using the same protocols and sites was carried out in 2008 to provide comparable data. This comparison, based within a relatively stable landscape, was intended to assess whether the continent-wide declines in grassland birds would also be evident at a local scale. Shrubland birds were also included in the monitoring, in part because grassland and low shrubland habitats often intergrade in Carden, and in part to provide a point of comparison with another guild of birds. Other habitats, such as forests and wetlands, were not monitored. The following 30 bird species were identified as target species for field monitoring because of their importance as grassland/shrubland species: American Goldfinch Eastern Meadowlark Northern Harrier American Kestral Eastern Towhee Northern Mockingbird Black-billed Cuckoo Field Sparrow Red-headed Woodpecker Bobolink Golden-winged Warbler Red-tailed Hawk Brown Thrasher Grasshopper Sparrow Savannah Sparrow Chestnut-sided Warbler Gray Catbird Short-eared Owl Clay-coloured Sparrow Horned Lark Song Sparrow Common Nighthawk Indigo Bunting Upland Sandpiper Eastern Bluebird Loggerhead Shrike Vesper Sparrow Eastern Kingbird Northern Flicker Whip-poor-will On the Carden Plain, some species require relatively large grassland or shrubland habitats for nesting. Because of their sensitivity to changes in habitat amounts, these species require special consideration in conservation efforts. The term area-sensitive describes a species whose presence or absence is closely related to the amount of breeding habitat area within a given spatial unit (Couturier, 1999). Among the selected IBA target species, the following 11 have been identified as being area-sensitive: Bobolink Horned Lark Savannah Sparrow Eastern Bluebird Loggerhead Shrike Short-eared Owl Eastern Meadowlark Northern Harrier Upland Sandpiper Grasshopper Sparrow Red-tailed Hawk 7

8 The selection of monitoring sites was structured to represent distinct habitat patches: areas of land differentiated by changes in habitat and cultural boundaries. A total of 93 habitat patches were sampled in both the 2000 and 2008 studies. As shown on Figure 2, these represent five different habitat types, including seven areas of Limestone Pavement, 49 Open Grassland, 20 Treed Grassland, 15 Shrubland and two Cropland habitat types. 1. Class I habitats, or limestone pavement, are characterized by patches of bare bedrock, often very dry in summer, with sparse vegetation. These habitats are infrequent and occur largely in the northern sections of the IBA. On these barren habitats, the most frequent species encountered include Savannah Sparrow, Eastern Meadowlark, Upland Sandpiper, and Horned Lark. These habitats are likely also important for nesting Common Nighthawk and Whip-poor-will along their edges. Vegetation characteristic of this habitat type include lichens, algae, mosses, Stonecrop, Early Saxifrage, Silvery Cinquefoil, Field Chickweed, and Stitchwort. Sixteen out of thirty IBA target species were found in Class I habitat, which comprised 610 hectares of the total IBA. 2. Class II habitats, or open grasslands, are much more abundant, and are characterized by moderately dense grasses and other forbs, with only a few scattered trees or shrubs. Most of these grasslands are pastured by cattle. Savannah Sparrow, Bobolink, Eastern Meadowlark and Upland Sandpiper are the most common birds in this habitat type, and it also supports the highest density of the less common Grasshopper Sparrow. Vegetation characteristic to this habitat type include Prairie Smoke, Harebell, Hairy Beardtongue, Mullein, and Hawthorn shrubs. Twenty-four target species were found in Class II habitat, which comprised 6530 hectares of the total IBA. 3. Class III habitats, or treed grasslands, are also generally pastured, but include more than 25% cover of trees and tall shrubs. Again, Savannah Sparrow and Eastern Meadowlark are the most frequently encountered species, with other common birds including American Goldfinch, Brown Thrasher, and Field Sparrow. The highest density of Loggerhead Shrike occurrences are also in this class. Characteristic vegetation includes a high ratio of grasses, clover and vetches, with smaller patches of wildflowers, and woody shrubs and small trees such as Fragrant Sumac, White Elm, White Birch, and Hawthorn. Twenty-two target species were found in Class III habitat, which comprised 1560 hectares of the total IBA. 4. Class IV habitats, or shrublands, are distinctly different, with relatively dense shrub cover amid a patchwork of grasslands and scattered trees. These habitats are often transitional, occurring on abandoned pasture lands, but they may also persist for long periods. Several different types are present, including those dominated by Hawthorns, Fragrant sumac, and Common juniper. The most common shrubland birds are Brown Thrasher, American Goldfinch, and Song Sparrow, but this habitat also supports the highest densities of such species as Eastern Towhee, Indigo Bunting, Clay-colored Sparrow, Vesper Sparrow, Gray Catbird, and Golden-winged Warbler. Vegetation characteristic of shrublands include Spreading Juniper, Fragrant and Staghorn Sumac, Gray Dogwood, Hawthorn, and Downy Arrowwood. Twenty-four target species were found in Class IV habitat, which comprised 3350 hectares of the total IBA. 8

9 5. Class V habitats, or croplands, include hayfields and more intensively managed rotational pasturelands, with dense uniform grass cover. Relatively few of these habitats were surveyed, but these areas appear to be particularly attractive to Bobolink and Eastern Meadowlark, with Eastern Kingbird also frequently occurring along field edges. While the density of birds in these productive areas is relatively high, only a few of the target species nest in this habitat type. Nine of thirty target species were found in Class V habitat, which comprised 2615 hectares of the IBA. During both the 2000 and 2008 breeding bird seasons, each monitoring plot was surveyed for the presence of birds at least twice. All counts were complete between May 31 and June 15, and between 5:00 and 11:00 in the morning. The semi-circular plots had a radius of one hundred meters delineated by estimation, giving each plot a sample size of 1.57 hectares. The number and species of birds seen or heard within the semi-circle during a 5-minute period were counted. In the 2000 survey, all field work was completed by a contractor, Terry Bradt, who visited many sites multiple times. In 2008, the task of counting the birds was given to teams of expert volunteers from the Ontario Field Ornithologists. Each volunteer team surveyed a group of approximately ten sites, with two five-minute counts carried out, separated by a two-minute quiet period. Synopsis of results: Because the original worksheets from the year 2000 monitoring had been misplaced, only a partial comparison of data was possible. In addition, because population levels may vary in response to short-term factors such as weather conditions, a comparison of only two points in time has to be treated with caution. Despite these limitations, several clear trends can be determined from the data comparison: Significant declines in the diversity of the 30 target species continue to occur at a rate of 2% - 4% per year, with shrubland birds showing the greatest decline (see Table 2). Open grassland habitats show the highest density of target bird populations in 2008, followed by tree grassland habitats (see Table 3). Area-sensitive species now occur much more frequently in open grassland habitats than other habitat types (see Table 4). Changes in abundance of individual species show a pronounced trend towards losses of short-grass specialists such as Horned Lark, and increases in a small group of shrubland species, notably Song Sparrow and Eastern Towhee (see Figure 3). 9

10 Figure 2: Habitat Distribution in the Carden Plain IBA igure 1: Classified alvar, grassland, and shrubland habitat within the Carden Plain IBA 10

11 Table 2: Change in Average Number of Target Species per Point Count between 2000 and 2008 Habitat Type # of Point Year 2000 Year 2008 Trend Counts Limestone pavement %/year decline Open grassland %/year decline Treed grassland %/year decline Shrubland % per year decline Table 3: Comparison of Density (Total # of Target Birds per Point Count) in 2008 Habitat Type # of Point Year 2000 Year 2008 Counts Limestone pavement 7 n/a 4.7 Open grassland 44 n/a 9.4 Treed grassland 19 n/a 6.8 Shrubland 12 n/a 5.0 Table 4: Comparison of Occurrence of Area-Sensitive Species in 2000 and 2008 Habitat Type # of Point Year 2000 Year 2008 Counts Limestone pavement Open grassland Treed grassland Shrubland Figure 3: Changes in Species Abundance Change in average number of individuals observed in 86 sample plots between 2000 and % 200% 150% 100% 50% 0% -50% -100% -150% HOLA AMKE UPSA EABL YSFL FISP VESP SASP GRSP BOBO EAME EAKI AMGO BRTH SOSP EATO change in % -84% -33% -15% -100% -7% -66% -9% 105% 40% 48% 34% 8% 36% 231% 162% species 11

12 B. Cameron Ranch and Windmill Ranch Annual Monitoring As a baseline for long-term monitoring of bird populations, a series of 16 permanent monitoring plots were established on Cameron Ranch in 2005, and an additional 13 plots established on Windmill Ranch in Each plot has a 100 metre radius. These two ranches are slated to be regulated as the Carden Alvar Provincial Park. They are both grazed by cattle, and this activity is intended to continue as a management tool for bird habitats. Figure 4: Monitoring plots on Cameron Ranch 12

13 Figure 5: Windmill Ranch Monitoring Stations Each spring during late May and June, volunteers from the Ontario Field Ornithologists and the Toronto Ornithological Club have completed two five-minute point counts at each monitoring station. A summary of monitoring results for the first five years has been compiled by the Couchiching Conservancy (Howes 2009). 13

14 Monitoring results: Based on the monitoring results to date, a total of 15 grassland and shrubland species were assessed, although three of these species (Horned Lark, Eastern Bluebird, and Golden-winged Warbler) were not present on both properties. While the length of the monitoring period is not sufficient to establish longterm trend lines, several observations raise interesting points for further study: The overall diversity of birds seen during monitoring visits is declining from 85 species in 2006 to 74 species in 2007, 71 species in 2008, and 63 species in Among the 15 species assessed, the numbers of individuals observed appears to be declining for 9 species (60%)on Cameron Ranch, and 5 species (33%) on Windmill Ranch, with a few other species showing increases (see Table 4). Table 5: Short-term population trends in assessed species Species Cameron Ranch Windmill Ranch Upland Sandpiper Increasing Decreasing Bobolink Stable Decreasing? Eastern Meadowlark Slightly increasing Decreasing Horned Lark Decreasing Sparse Eastern Loggerhead Shrike Decreasing Decreasing Grasshopper Sparrow Increasing Increasing Vesper Sparrow Increasing? Decreasing Savannah Sparrow Increasing? Strongly decreasing Northern Harrier Sparse Sparse Henslow s Sparrow No data Occasionally present but sparse Short-eared Owl No data No data Eastern Bluebird No data Strongly decreasing Brown Thrasher Slightly decreasing Slightly decreasing Eastern Towhee Increasing Decreasing Golden-winged Warbler No data Decreasing Clay-colored Sparrow Increasing Stable Black-billed Cuckoo Increasing Sparse Grey Catbird Decreasing Strongly decreasing Brown-headed Cowbird Decreasing Strongly decreasing Perhaps the most surprising aspects of these results are: the frequent disparity in trends observed between the two ranches, the 5 out of 7 shrubland species found to be declining on Windmill Ranch, the increasing frequency of Grasshopper Sparrows. Ongoing monitoring will continue to help identify the factors at work influencing these trends. 14

15 C. Monitoring Plots in Ramara Township In response to recommendations in the Carden Conservation Plan, a volunteer-based survey of grassland birds within the Ramara section of the Carden Plain was carried out in the spring of In advance of the survey, habitat patches were mapped using the same criteria as in the year 2000 and 2008 Carden IBA programs. No limestone pavement sites were mapped in Ramara, but the project mapped 26 open grassland, 7 treed grassland, 7 shrubland, and 19 cropland (hay and managed pasture) sites. A point count monitoring plot was selected to represent each habitat patch, with the plots clustered in eight areas as shown in Figure 6. Volunteers carried out two point counts at each plot, using a monitoring protocol identical to that used in the Carden IBA. The intent of this project was to assess how grassland and shrubland habitats in Ramara compare to their counterparts just to the east. Figure 6: Location of monitoring plots in Ramara 15

16 Monitoring results: A compilation of monitoring results from the 59 Ramara monitoring plots is shown in the tables below, together with a comparison of equivalent results from the CKL North and CKL South sections of the Carden Plain (Figure 1). Table 6: Comparison of Average Number of Target Species per Point Count Habitat Type Ramara CKL North CKL South Cropland (Hay) 3.0 n/a n/a Open grassland Treed grassland Shrubland n/a Table 7: Comparison of Density (Total # of Target Birds per Point Count) Habitat Type Ramara CKL North CKL South Cropland (Hay) 3.5 n/a n/a Open grassland Treed grassland Shrubland n/a Table 8: Comparison of Occurrence of Area-Sensitive Species (# per point count) Habitat Type Ramara CKL North CKL South Cropland (Hay) 1.4 n/a n/a Open grassland Treed grassland Shrubland n/a As Tables 6-8 illustrate, habitat patches in Ramara consistently rank lower than either CKL North or CKL South in all of the habitat classes, and in comparisons of the number of species present, their density, and the number of area-sensitive species observed. The reasons for this disparity are not entirely clear. One factor which may be significant is the generally smaller size of habitat patches in Ramara, and the frequent presence of wooded fencerows that cause greater fragmentation, especially of open grasslands. Habitats in Ramara also appear to be more scattered, with longer stretches of woodland or agricultural croplands between patches, compared to the CKL sites. It is also possible that habitats in Ramara are more intensively managed for farming, although this requires further study. A more detailed comparison of area-sensitive grassland birds: Five species of grassland birds which occur widely on the Carden Plain exhibit characteristics of areasensitivity i.e. they require relatively large patches of grassland habitat to successfully breed. As shown in Figure 7, Ramara hayfields, and to a lesser extent Ramara grasslands, are consistently less welcoming for area-sensitive species. This pattern is especially pronounced for Upland Sandpiper and Eastern Meadowlark. The other conclusion to be drawn from this graph is the importance of the extensive grasslands of north Eldon Township (CKL South) to this group of bird species. 16

17 Figure 7: Comparison of Area-sensitive Grassland Species Occurrence of Area-Sensitive Grassland Species as a Percentage of Sites Surveyed Eastern Meadowlark Bobolink Grasshopper Sparrow Savannah Sparrow Upland Sandpiper Ramara Hay Ramara Grassland CKL South CKL North % of point counts CKL North CKL South Ramara Grassland Ramara Hay Upland Sandpiper 52% 70% 11% 0 Savannah Sparrow 61% 100% 69% 63% Grasshopper Sparrow 32% 40% 11% 5% Bobolink 70% 80% 76% 57% Eastern Meadowlark 85% 100% 53% 15% Significant Grassland Areas in Ramara: As a final factor in comparing Ramara grasslands to those in the Carden IBA, individual habitat patches were assessed against a modified set of significance criteria developed through the Carden IBA Conservation Action Plan (Coxon and Reid 2001). As a threshold of significance, these criteria identify as significant any habitat patch with a point count plot containing: 7 or more target species 5 or more area-sensitive species, or 15 or more birds among all target species combined. No patches of Ramara shrubland or cropland meet these criteria, but a single patch of treed grassland (MA0606) and three patches of grassland (MA 0104, MA0207, DA0108) do qualify. The latter site is actually located in former Dalton Township north of Lake Dalrymple, and the treed grassland site is just east of the southern end of the lake in the Lakeview Beach area. The other two grassland sites are located south and east of the Lafarge quarry south of Brechin, as shown on Figure 8. These sites are close to other significant grassland areas just to the east that were identified as significant during the earlier Carden IBA surveys. Similarly, areas of limestone pavement and grassland on both sides of Simcoe Rd. 47 along Ramara s eastern boundary had been previously assessed as significant. 17

18 Figure 8: Grassland Habitat Patches South of Brechin Note that habitat patches MA0104 and MA0207 ranked as significant grasslands, comparable to areas along Concession 1 just to the east which were included in earlier Carden IBA monitoring. 18

19 Conclusions 1. Birds of the Carden Plain are not exempt from the documented regional declines in grassland species, with average annual declines of 2% to 4% over the past decade, even though habitat conditions and extent do not appear to have changed significantly. 2. Open grassland habitats support the highest density of target bird species, and have maintained the highest frequency of area-sensitive species. 3. Changes in abundance of individual species show a pronounced trend towards losses of short-grass specialists such as Horned Lark, and increases in a small group of shrubland species, notably Song Sparrow and Eastern Towhee. 4. Diversity of shrubland species is showing a strong decline, at the same time as the abundance of two shrubland species is increasing rapidly. 5. Annual monitoring shows declines for 9 grassland species and 5 of 7 shrubland species, but with considerable unexplained variation between Cameron Ranch and Windmill Ranch plots. 6. Habitat patches in Ramara Township consistently rank lower in target bird diversity and density than the Carden IBA plots to the east, with that pattern especially pronounced for Upland Sandpiper and Eastern Meadowlark, both area-sensitive species. 7. The most significant grassland areas identified in Ramara are clustered southeast of Brechin and north of the Trent-Severn Canal. 8. The extensive grasslands of the CKL South section (north half of former Eldon Township) stand out as particularly diverse and productive in their bird life. 19

20 Literature Cited Bird Studies Canada Grassland Bird Conservation Action Planning in Southern Ontario: Workshop Backgrounder & Discussion Document. Bird Studies Canada, Port Rowan, Ontario. 14 pp. Cadman, Michael D., Donald A. Sutherland, Gregor G. Beck, Denis LePage, Andrew R. Couturier, eds Atlas of the Breeding Birds of Ontario: Bird Studies Canada, Environment Canada, Ontario Field Ornithologists, Ontario Ministry of Natural Resources, Ontario Nature. Chapman, L.J. and D.F. Putnam The Physiography of Southern Ontario. Ontario Ministry of Natural Resources, Toronto. Couchiching Conservancy, The Towards an Integrated Carden Conservation Strategy. Part 1: Healthy Ecosystems & Species at Risk. Orillia, 42 pp. Couturier, A Conservation Priorities for the Birds of Southern Ontario. Unpublished Bird Studies Canada report. 17 pp. Coxon, A. and R. Reid Carden Plain Important Bird Area: Conservation Action Plan. The Carden Plain IBA Steering Committee and The Couchiching Conservancy. Orillia. 73 pp. Eco Source Management Carden Conservation Plan Working Draft Technical Report. The Couchiching Conservancy, Orillia. 43 pp. Heagy, Audrey E. and J.D. McCracken Monitoring the State of Ontario Migratory Land Birds. Bird Studies Canada, Port Rowan, Ontario. Howes, Kyra Upland Birds of Carden Plain: Species Summary and Habitat Recommendations for Grassland Management. The Couchiching Conservancy, Orillia. 70 pp. McCracken, J.D Where the Bobolinks Roam: The Plight of North America s Grassland Birds. Bird Studies Canada, Port Rowan, Ontario. Pashley, D.N., C.J. Beardmore, J.A. Fitzgerald, R.P. Ford, W.C. Hunter, M.S. Morrison, K.V. Rosenberg Partners in Flight, Conservation of the Land Birds of the United States. American Bird Conservancy, Virginia. Sauer, J.R., J.E. Hines, and J. Fallon The North American Breeding Bird Survey, results and analysis, USGS Patuxent Wildlife Research Center, Laurel. 20