MMP Protocol. Route and Station Selection and Characteristics. Amphibian Survey Protocol. What is marsh habitat?

Transcription

1 MMP Protocol Route and Station Selection and Characteristics Upon registering with the MMP, volunteers receive an MMP Training Kit that includes: a detailed protocol booklet; field and summary data forms; an instructional cassette tape with examples of the songs and calls of the birds and amphibians most likely to be encountered in the Great Lakes basin; and a broadcast tape used to elicit calls from some of the more secretive wetland birds. MMP participants also receive an annual newsletter that summarizes survey results and includes articles on the conservation and ecology of wetlands, amphibians and birds. Although participants are free to establish survey routes in wetlands of their choice, MMP staff work with new volunteers to help ensure MMP routes do not overlap and that routes are established in wetlands in need of survey coverage. Each MMP route may consist of from one to eight stations but the route must be able to be surveyed in its entirety by a single observer on a single evening. Stations surveyed for birds must be separated by at least 250 metres (275 yards) to minimize duplicate counts of the same individual birds. Because judging distances and the locations of calling amphibians can be difficult, stations surveyed for amphibians must be separated by at least 500 metres (550 yards). MMP staff encourage surveys for both amphibians and birds on as many routes as possible. MMP Volunteer: D. Tozer An MMP station is defined as a 100 metre (110 yard) radius semicircle with more than 50% of the area within the semicircle covered by marsh habitat (see box What is marsh habitat? ). Counts are conducted from a focal point at each station the surveyor stands at the midpoint of the 200 metre (220 yard) base of the semi-circle and faces the arc of the station perimeter. Each focal point is permanently marked with a stake and metal tag to facilitate relocation in subsequent years. MMP stations are usually accessed along the edge of marshes but volunteers interested in monitoring stations accessible only by boat or canoe are encouraged to do so. Northern Leopard Frog: BSC Files What is marsh habitat? MMP stations must be dominated by marsh habitat. Such habitat typically occurs in wetlands that are regularly or periodically wet or flooded to a depth of up to two metres (six feet). Although shrubs and trees may be present, non-woody, emergent plants such as cattail, bulrush and Burreed are predominant. These emergent plants often occur intermixed with shallow open water areas, usually containing submerged and floating plants such as bladderwort and duckweed. Marsh habitat may occur along river edges, in lake bays, or mingled with dead standing timber. Thus far, the MMP has focussed on marsh habitat because marshes are one of the most productive wetland types in the world and are home to a wide range of wetland-dependent amphibian and bird species. Amphibian Survey Protocol Amphibians surveyed by the MMP protocol are those frogs and toads that typically depend on marsh habitat for their spring and early summer mating rituals. Males of several frog and toad species announce their presence to females through distinctive calls, grunts and chirps. MMP routes are surveyed for these calling amphibians on three nights each year, with at least 15 days between visits. Because peak amphibian calling periods are closely linked to temperature and other weather conditions, visits are scheduled according to night-time air temperatures rather than dates. The MMP protocol encourages volunteers to conduct their first visit soon after night-time air temperatures reach 5 C (41 F) and early spring rain showers. Second and third visits should occur soon after night-time temperatures reach 10 C (50 F) and 17 C (63 F), respectively. Amphibian surveys are to begin one-half hour after sunset and end before midnight. Visits are to be conducted on evenings with little wind, preferably under warm, moist air conditions. During the 3-minute survey visit, observers assign a Call Level Code to each species detected and estimate numbers of individuals for two of these levels. A Call Level Code of 1 is assigned if calls of individual frogs or toads do not overlap and calling individuals can be counted. Code 2 is assigned if the calls of individuals sometimes overlap but the number of individuals can reasonably be estimated. Code 3 is assigned if so many individuals of a species are calling that overlap among calls seems continuous; a count estimate is impossible for Code 3 and is not required by the protocol. Judging distances accurately can be particularly difficult in the dark, and through 1998, surveyors were not asked to differentiate between amphibians calling from within and outside the station boundary. Starting in 1999, MMP participants were asked to use their best judgement to distinguish whether each species detected was calling from inside the station boundary only, from outside the boundary only, or from both inside and outside. Combined with habitat information provided for each station by MMP surveyors, this modification will help provide better information on amphibian habitat associations. 10

2 MMP Protocol Bird Survey Protocol Survey visits for birds are conducted twice each year, at least 10 days apart, between May 20 and July 5. Visits must begin after 6:00 p.m. and are to be conducted in good survey conditions (i.e. warm, dry weather with little wind). A 5-minute broadcast tape is played at each station during the first half of each 10-minute survey visit. The broadcast tape helps elicit calls from several normally elusive species and contains the calls of Virginia Rail, Sora, Least Bittern, Common Moorhen, American Coot and Pied-billed Grebe. During the count period, observers record all birds heard or seen within the station area and map their observations onto a field sheet. Aerial foragers are also counted and are defined as those birds actively foraging within the station area to a height of 100 metres (110 yards). Bird species flying through or detected outside the station area are tallied separately. Pied-billed Grebe I. Jeklin/Cornell Laboratory of Ornithology Habitat Assessments Habitat associations for many species of Great Lakes wetland birds and amphibians are not well known. A good understanding of these relationships is critical to designing effective wetland management and conservation practices. When combined with information on trends in species occurrence or abundance, data on vegetation and other wetland characteristics help identify those wetland habitats most at risk of losing their ability to support marsh birds and amphibians. Using a standard protocol, surveyors are asked to make an annual assessment of the habitat characteristics for each of their MMP stations. Volunteers are encouraged to conduct the habitat surveys in late May to mid-june, when the plants can be readily identified. Observers provide information on the coverage of five general habitat types: herbaceous emergent plants; open water; exposed mud, rock or sand; trees; and shrubs. The percent coverage of the four most dominant types of emergent plants is also recorded, providing a more detailed assessment of this important component of wetland habitat. Observers record the coverage of floating plants and provide their best assessments of wetland size and permanency, and adjacent land use. Observers also make a simple sketch map of the general habitat characteristics of each station. Water Lily: BSC Files 11

3 The summaries of data presented in this report are intended as an overview of the types of information contributed by MMP surveyors and to demonstrate the breadth of future analyses. Although the MMP has been successful in compiling more than five years of survey data from a large geographic area, the program is still quite young for a population-monitoring program. As demonstrated in this section, additional years of data will lead to improved resolution on trends for both amphibians and birds. Detailed habitat, trend and other analyses are currently possible but must be customized to address particular questions at specific geographic and time scales. These analyses are beyond the scope of this report but are being pursued as sufficient data accumulates and as specific information needs are identified. Routes In total, 493 different participants submitted data from 575 routes during the period 1995 through Most routes (489 routes, 85% of total) were within the Great Lakes basin (Figure 2). The Lake Erie basin contained the most routes (156, 32% of total), followed by the Ontario (139, 28%) and Huron (102, 21%) basins; fewer routes were in the Michigan (68, 14%) and Superior (24, 5%) basins (Table 1). Fewer MMP routes have been established in the Lake Superior basin than in other basins due to the relative scarcity of both appropriate marsh habitats and available surveyors in the Superior region. Within the Great Lakes basin, amphibian survey data from 354 routes were submitted and bird survey information was provided from 322 routes. Similar numbers of routes were surveyed for amphibians only, birds only and for both groups (Table 1). The number of routes for which data were submitted averaged about 230 routes per year and peaked in 1997 (Table 1). Gains were made in the U.S. from 1995 to In 1995, the number of routes in the U.S. (65 in total) was about half of the total routes in Canada that year (120). However, over the first five years of the MMP the number of U.S. routes has grown to 99, comparable to the number of routes currently surveyed in Canada (116 routes). Figure 2: Locations of MMP routes inside and outside the Great Lakes basin. Routes surveyed at least once 1995 through 1999 are indicated by black circles. Table 1: The number of routes surveyed in each lake basin, summarized for routes monitored for amphibians (a), birds (b) and both groups (ab), 1995 through

4 Of the routes surveyed for amphibians at least once in the period 1995 through 1999, 32% were established in 1995, 30% in 1996, 25% in 1997 and the remainder started in 1998 or Overall, a large percentage (44%) of amphibian routes were surveyed for 1 year only, fewer for 2 or 3 years (19% each) and smaller percentages for 4 and 5 years (14% and 4%, respectively) (Table 2). Of the routes surveyed for birds at least once in the period 1995 through 1999, 44% were established in 1995, 18% in 1996, 6% in 1997 and the remaining 5% in As for amphibian routes, a large percentage (43%) of bird routes were surveyed for 1 year only. Eighteen percent of bird routes were monitored for 2 years, and 12% each for 3 and 4 years. A higher proportion of bird routes were monitored for the full 5-year period (15%), than for amphibian routes (i.e. 4%). Routes within the Great Lakes basin had an average of 3.9 stations. The average number of stations on routes was similar for bird and amphibian routes and varied little across years or among lake basins (Figure 3). Table 2: The number and percent of amphibian and bird routes surveyed for 1 to 5 years, 1995 through A B Figure 3: Average number of survey stations on MMP routes surveyed for amphibians and birds, summarized by A) year and B) lake basin. Black vertical bars are 95% confidence limits. I What are confidence limits? Surveys conducted under the MMP s monitoring protocol are samples rather than complete counts. Because the true values remain unknown, there is some uncertainty associated with trends and other values derived from sampling. This uncertainty can be represented by 95% confidence limits, which indicate that the true value is thought (with 95% certainty) to lie between the upper and lower extremes of the confidence limit. Narrow (or wide) confidence limits indicate that the uncertainty associated with an estimated trend or other average value is low (or high). Trend estimates with confidence limits that do not include zero indicate significant change with time. I Please see text box 'What are Confidence Limits?'. 13

5 Overall, most (76.7% of station-years) MMP surveys were conducted in permanent wetlands with the rest done in semi-permanent (15.8%) and seasonal (7.5%) wetlands. There was an increase in the proportion of stations in semi-permanent wetlands after 1996 (Figure 4a). Lake Superior had the greatest proportion of stations in permanent wetlands while a smaller percentage of Lake Michigan s stations were in this wetland class (Figure 4b). Overall, approximately 32% of stations were in huge wetlands, 19% in large wetlands, a similar proportion (20%) in tiny wetlands and the remainder split about evenly between medium and small wetlands (15% and 14%, respectively) (see Glossary for size definitions). This distribution of stations among wetland sizes fluctuated somewhat among years (Figure 5a). Among lake basins, only stations in Lake Michigan were approximately evenly distributed among wetland size classes. The huge wetland size class predominated for Lakes Ontario, Erie and Huron while huge and large wetland size classes predominated in Lake Superior (Figure 5b). What are route-years and station-years? Analyses that considered every route surveyed in a given year as a single observation and did not differentiate between routes surveyed for single and multiple years were summaries of MMP data in terms of route-years. Similarly, the term station-year refers to the basic sample unit for those analyses that considered stations without regard to the number of years each station was surveyed. Unless otherwise mentioned, most analyses in this report were based on route-year and station-year approaches. A B Figure 4: The percent of MMP stations surveyed on seasonal, semi-permanent and permanent wetlands for each A) year and B) lake basin, 1995 through A B Figure 5: The percent of MMP stations surveyed on huge, large, medium, small and tiny wetlands for each A) year and B) lake basin, 1995 through

6 Habitat Because some details of the MMP habitat assessment protocols were revised prior to the 1997 survey season, habitat characteristics are summarized in this report for 1997 onward only. Overall, stations within the Great Lakes basin had an average of 55.5% of the station area covered with non-woody, emergent vegetation and an average of 26.9% of the area in open water. Exposed substrates (e.g. mud, rock), trees and shrubs averaged much lower coverage at 2.5, 7.6 and 7.4% respectively. These patterns were observed in all lake basins, although stations in Lake Superior had higher average coverage of open water than stations in other basins (Figure 6). Plant species contributing more than 10% to the overall coverage of non-woody, emergent vegetation were considered dominant; all others were classified as subdominant. Cattail was the most frequent dominant emergent plant at stations in the Great Lakes basin (44.4%), but grasses/sedges were also dominant at 21% of stations. A variety of other plant species dominated 19% of stations and a diverse mix of species, none dominant, occurred at 15.6% of stations. Some variation among basins was observed (Figure 7). Cattail was a particularly strong element of the emergent vegetation in Lake Ontario stations. Cattail and grasses/sedges were dominant in about equal proportion of Lake Huron stations. Cattail made up only a small part of the emergent cover in Lake Superior stations; most of these stations were dominated by either grasses/sedges or were characterized by a mix of subdominant species. Cattail and grasses/sedges were present in most station-years (about 80% and 60%, respectively) (Figure 8a). Most other emergent plants were recorded at less than 20% of stations. Where they did occur, cattail and grass/sedge tended to dominate, usually comprising more than 19% of the emergent plant community (Figure 8b). Where species other than cattail and grasses/sedges occurred, most tended to make relatively minor contributions (1 to 19%) to the overall emergent plant cover for most stations. However, where they did occur, Wild Rice, rushes, Smartweed, Purple Loosestrife, Water Willow and Common Reed each contributed at least 20% cover to the emergent community of a substantial percentage (at least 40%) of stations. Figure 6: The average percent of station area in each lake basin covered by emergent vegetation, open water, exposed substrate, trees, and shrubs, 1997 through Figure 7: The average percent of station area in each lake basin (1997 through 1999) for which the dominant emergent vegetation was cattail, grass/sedge, other emergent plant species, or a mix of subdominant species. A B Figure 8: A) The percent of station-years with each emergent plant type, 1997 through B) At those stations where each plant type occurred (i.e. more than 0% coverage), the proportion of station-years where the plant was recorded in each of four possible coverage classes, 1997 through For each plant type, coverage classes represent the percent of the total emergent plant community composed of that plant type (e.g. where grass/sedge occurred, 1-19% of the emergent plant community was grass/sedge for about 30% of station-years). 15

7 Amphibians MMP surveyors recorded a total of 14 species of calling amphibians during the 1995 through 1999 period. All but one species, the Great Plains Toad, occurred on MMP routes within the Great Lakes basin (Table 3). This report focuses on the results of MMP surveys within the Great Lakes basin and emphasizes those calling amphibians most closely associated with wetlands and other aquatic habitats. Bullfrog and Green Frog are the most aquatic, spending the winter in wetland sediments and often requiring two summers as tadpoles before their metamorphosis into the adult stage (Tyning, 1990). The Northern Leopard Frog also breeds in aquatic habitats and hibernates in wetland sediments but, because it completes the tadpole stage in a single season and is frequently found in terrestrial habitats outside peak breeding periods, this species is not considered as strongly aquatic. Grey Treefrog, Chorus Frog and Spring Peeper are somewhat Pickerel Frog: BSC Files less closely associated with aquatic systems. All 3 species breed on wetland edges but Chorus Frog and Spring Peeper also make use of temporary, shallow habitats such as roadside ditches and flooded fields. All three are also associated with treed and shrubby habitats, and over winter in upland areas under logs, leaf litter and loose bark. American and Fowler s toads are the most terrestrial species recorded by MMP surveyors, depending on wetlands and temporary pools for breeding but spending most of their time in upland habitats. Spring Peeper was the most frequently detected species (Table 3) and, as indicated by an average calling code of 2.5, was frequently recorded in full chorus (Call Level Code 3) where it was encountered. Green Frog was detected in more than half of station-years and the average calling code indicates this species was usually recorded as Call Level 1. Grey Treefrog, American Toad and Northern Leopard Frog were also common, being recorded in more than one-third of all station-years. Grey Treefrog was recorded with the second highest average calling code (1.9), indicating that MMP observers usually heard several individuals with some overlapping calls. Bullfrog, Chorus Frog and Wood Frog were detected in approximately one-quarter of station-years. Five species were detected infrequently by MMP surveyors and were recorded in less than 3% of station-years. Some of these latter species are quite rare in parts of the Great Lakes basin and are the subject of monitoring efforts more intensive than the MMP. Because records of rare species have not been rigorously verified by MMP staff, this report is focused on the more common species. Because the relationship between calling codes and numbers of individuals is not well known, the focus of this report is on amphibian species presence (or occurrence) as measured by MMP surveyors. The 8 amphibian species commonly detected (present in at least 3% of station-years) by MMP surveyors varied somewhat in their distribution among lake basins (Figure 9). Because the range of each species extends the breadth of the Great Lakes basin, these patterns are not likely due to range limitations. Differences in habitats, regional population densities, timing of survey visits or other factors are more likely explanations. Northern Leopard Frog was detected most frequently in the Ontario and Erie basins. Green Frog occurred with about equal frequency in the Lake Ontario, Erie, Huron and Michigan basins while the Bullfrog was recorded most frequently in the Ontario and Erie basins. Spring Peeper was encountered frequently in the Huron, Michigan and Superior basins and occurred less often in the two southernmost lake basins. Figure 9: Percent of each lake basin s station-years in which each of the common amphibian species was present on Great Lakes MMP routes, 1995 through Species are ordered by taxonomic similarity. 16 Table 3: Frequency of occurrence and average calling code for amphibian species detected inside Great Lakes basin MMP stations, 1995 through Amphibian data were collected from 2,935 station-years. Species are ordered by decreasing frequency of occurrence.

8 Trends in station occupancy were assessed for the 8 amphibian species commonly detected on Great Lakes MMP routes. Because natural variability can be substantial and station occupancy is a relatively coarse measure (i.e. does not detect changing numbers), trend estimates for amphibians should be regarded as preliminary. For each species, a trend was assessed first on a route-by-route basis in terms of the annual proportion of stations with each species present. These route level trends were then combined for an overall assessment of trend for each species. Although some trends were suggested for species such as American Toad and Bullfrog, only the declining trend for Chorus Frog could be resolved with sufficient statistical confidence (i.e. confidence limits do not encompass zero) (Figure 10). Declines in Chorus Frog station occupancy appeared to be driven by declines in the Huron and Michigan lake basins; few stations were occupied in the Ontario basin and Chorus Frog presence fluctuated but did exhibit a trend in the Erie basin (Figure 11). Although long-term (1950s to 1990s) losses of Chorus Frog have been recorded in the St. Lawrence River valley just outside the Great Lakes basin, this species is known to have population fluctuations, and even regional extinctions, over short time periods due to natural factors such as differences in annual weather conditions (Diagle, 1997). Although widespread declines in some regional populations (e.g. western Northern Leopard Frog populations) are feared to be irreversible, anecdotal and research evidence suggests that wide variations in the occurrence of many amphibian species at a given site is a natural and ongoing phenomenon (Green, 1997). Figure 10: Annual indices of calling amphibian occurrence on MMP routes within the Great Lakes basin, 1995 to Indices are based on the annual proportion of survey stations with each species present and are defined relative to 1999 values; vertical bars indicate 95% confidence limits I around annual indices. The estimated annual percent change (trend) is indicated for each species and the associated lower and upper extremes of 95% confidence limits are enclosed in parentheses. Routes included in analyses were those inside the Great Lakes basin and with at least 3 annual survey visits and at least 2 years of survey coverage. Only species detected on at least 15 routes were included in trend analyses. Trend estimates are summaries of within-route trends as calculated with logistic regression. Figure 11: Annual indices for Chorus Frog, summarized for those lake basins in which it was detected on at least 15 routes, 1995 through I Please see text box 'What are Confidence Limits?' (page 13). 17

9 As figures 10 and 11 indicate, annual fluctuations are apparent for many of the species monitored by MMP surveyors. In addition to natural fluctuations in local populations, calling amphibians may be present at a monitoring station but not detected for a wide variety of reasons (e.g., temperature). Against this backdrop of variability, the five years of MMP survey data is a short timeframe and the low resolution on trends in amphibian station occupancy is therefore not surprising. Additional years of MMP monitoring data, particularly if augmented with intensive studies of individual species, will help reveal whether the patterns observed in the early stages of MMP monitoring continue. The extent to which additional years of data might be expected to provide good resolution on amphibian occupancy trends was assessed based on MMP data collected 1995 through The annual trend (i.e. percent change in population index based on station occupancy) that could be detected was estimated assuming either 100 or 200 routes were monitored over 11 years (a 10-year interval) (Table 4). With 100 routes measured for 11 years, the estimated annual trend that could be detected was about 1% per year or less for all of the eight amphibians commonly recorded on MMP routes. As expected, resolution improved with 200 routes measured for 11 years. Expected resolution on trends was best for American Toad and Table 4: Expected annual percent change in the proportion of stations with each amphibian species that are likely to be detected with 100 or 200 routes surveyed over a 10-year interval (11 years of monitoring). Green Frog, followed by Grey Treefrog and Northern Leopard Frog (Table 4). Resolution was lower for species that were less common (Bullfrog, Chorus Frog and Wood Frog) or which exhibited large fluctuations in station occupancy (Spring Peeper; see Figure 10). These results suggest that additional years of data are likely to improve the ability of MMP data to provide trend estimates for several amphibian species, including species of conservation and management concern such as the Northern Leopard Frog and Bullfrog. Detectable percent annual changes were based on binomial regression, alpha and beta set to 0.05 and 0.80 respectively, and were adjusted for the frequency of occurrence of each species in MMP surveys, 1995 through Wood Frog: BSC Files Grey Treefrog: BSC Files 18

10 Birds Overall, a total of 207 bird species was recorded by MMP observers from 1995 through 1999, with 191 of these species counted inside the standard MMP station boundary. In this report, MMP bird survey data are summarized for stations within the Great Lakes basin only and for bird species closely associated with wetlands those that use marshes for feeding, nesting or both. Fiftythree of these species were recorded by MMP observers on Great Lakes routes, with 45 classified as typically nesting in marshes (marsh nesters) and eight using the air space above marshes for feeding (aerial foragers; may also nest in wetlands). Although data are presented for American Coot and Common Moorhen, their calls can often be difficult to distinguish. Records for which these species were not differentiated by MMP surveyors are also summarized in this report as a distinct species. Tree and Barn swallows were the most commonly occurring aerial foragers, recorded in about half and onequarter of station-years respectively (Table 5). The other six aerial foraging species occurred much less frequently (<8% of station-years). Where they were observed, Tree and Bank swallows tended to occur in larger numbers than other aerial foragers (Table 5). Red-winged Blackbird was clearly the most commonly recorded marsh nesting species, occurring at about 88% of station-years. Swamp Sparrow was observed in almost half of all station-years, and four other songbirds (Common Yellowthroat, Yellow Warbler, Song Sparrow and Marsh Wren) were almost as common. Several species closely associated with marshes were observed in about 10 to 25% of stationyears. Of special note among these were several water birds not well surveyed by other monitoring programs: Virginia Rail, moorhen/coot (undifferentiated), Black Tern, Common Moorhen, Pied-billed Grebe and Sora. Where they were recorded, Red-winged Blackbird (only males are counted by MMP observers) occurred in the highest numbers at 5.4 per station, followed by Canada Goose with an average of 5.2 individuals per station. Common Grackle, moorhen/coot and Black Tern each averaged at least 4 individuals per station where they occurred. In contrast, very few stations with bitterns and Willow Flycatchers contained more than a single individual. MMP Results Virginia Rail: S. J. Lang/VIREO Table 5: Frequency of occurrence and average number of individuals for aerial forager and marsh nesting bird species detected inside Great Lakes basin MMP stations, 1995 through Bird data were collected from 3,003 station-years. Species are presented by group and ordered by decreasing frequency of occurrence. 19

11 The 28 bird species commonly recorded (present in at least 3% of station-years) by MMP surveyors varied considerably in their distribution among lake basins (Figure 12). Many of these patterns were primarily related to differences in species geographic range but may have also been influenced by the vegetation and other wetland habitat characteristics of the surveyed stations. MMP data might be expected to provide the best information on bird populations and wetland habitat when the geographic scale of analyses matches the regions of strongest species occurrence. In general, aerial foraging and marsh nesting birds tended to occur at the greatest proportion of station-years in the Lake Ontario basin (average proportion of station-years = 0.25) and at the lowest proportion of stations in the Lake Superior basin (0.14). Average values for the Lakes Erie and Huron (0.21 each) and Lake Michigan (0.19) basins were intermediate. This suggests that, in general, analysis of MMP bird species information will provide the greatest resolution for trends in the Ontario, Erie and Huron basins but may be more limited for the Michigan and Superior basins. Most marsh nesting and aerial foraging birds were detected more frequently on stations on routes in the four lower Great Lakes than on routes in the Lake Superior basin (Figure 12). However, several of the most commonly recorded bird species (Tree Swallow, Yellow Warbler, Common Yellowthroat, Song Sparrow, Swamp Sparrow, Red-winged Blackbird) were detected on a substantial proportion (at least 30%) of Lake Superior stations. Even so, this suggests that MMP data on many of the most common bird species have the potential for conveying basin-wide patterns and that MMP resolution may be more limited for assessing population and habitat trends in the Lake Superior basin. Although additional years of data are necessary to assess population trends on the level of individual lake basins, these assessments are likely to be strongest for those lake basins in which species occur most frequently. Some closely related species differed somewhat in the frequency of their occurrence. American Bittern was detected most frequently in the Lake Huron basin while Least Bittern occurred in similar proportions of station-years across the Ontario, Huron and Erie basins. Virginia Rail and Sora also tended to differ in their occurrence, with the former detected most often in the Ontario and Huron basins and the latter found in somewhat similar proportions of stations across all basins except Lake Erie. Almost all records of Alder Flycatcher were from the Lake Superior basin, while those of Willow Flycatcher were broadly distributed across the Ontario, Erie and Michigan basins with few records in the Huron and Superior basins. Most Pied-billed Grebes were detected in the Lake Huron basin with relatively few recorded in the Lake Superior basin. Black Tern was found most often in the Lake Huron basin and was not recorded in the Lake Superior basin. MMP Results With only five years of data collected across the Great Lakes basin, the MMP is still quite young as a populationmonitoring program. Bird species occurrence and numbers, and their activity and likelihood of being observed, vary naturally among years and within seasons. For these and other reasons, large numbers of observations, collected over many years, are required to estimate population trends reliably. Although results are Figure 12: Percent of each lake basin s station-years in which each of the common marsh nesting and aerial foraging birds was present on Great Lakes MMP routes, 1995 through Species are ordered by taxonomic similarity. 20

12 still preliminary, trends are presented for several birds recorded on Great Lakes MMP routes (Figure 13a,b). Population indices and trends (i.e. average annual percent change in population index) are presented for species with statistically significant trends, 1995 through Species with significant basin-wide declines were Pied-billed Grebe, Blue-winged Teal, American Coot, undifferentiated moorhen/coot, and Black Tern (Figure 13a). Although declines for Tree Swallow and Red-winged Blackbird are less certain, data for these species are also summarized in Figure 13a because these are particularly widespread and common marsh nesting birds and declines in their numbers merit some scrutiny. Statistically significant basin-wide increases were observed for Canada Goose, Mallard, Chimney Swift, Northern Rough-winged Swallow, Common Yellowthroat and Common Grackle (Figure 13b). Each of the declining species depends upon wetlands for breeding but, because they use wetland habitats almost exclusively, the Pied-billed Grebe, coot, moorhen, and Black Tern are particularly dependent on the availability of healthy wetlands. Although these declines in some wetland specialists and increases in some wetland edge (e.g. Common Yellowthroat) and generalist (e.g. Canada Goose) species suggest trends in wetland habitat conditions, additional years of data and more detailed analyses are required to understand how these patterns relate to trends in Great Lakes wetland function. A B Figure 13: Annual population indices of A) declining and B) increasing marsh nesting and aerial foraging bird species detected on Great Lakes basin MMP routes, 1995 through Population indices are based on counts of individuals inside the MMP station boundary and are defined relative to 1999 values; vertical bars indicate 95% confidence limits I around annual indices. The estimated annual percent change (trend) is indicated for each species and the associated lower and upper extremes of 95% confidence limits are enclosed in parentheses. Routes included in analyses were those within the Great Lakes basin and with at least 2 annual survey visits and at least 2 years of survey coverage. Only species detected on at least 15 routes were included in trend analyses. Trend estimates are summaries of within-route trends as measured by Poisson regression and are corrected for overdispersion. I Please see text box 'What are Confidence Limits?' (page 13). 21

13 For many species, changes in numbers were strongest in only some lake basins. For example, the decline in Pied-billed Grebe (Figure 14a) appeared to be largely due to trends in Lakes Ontario and Huron MMP routes; patterns in the other lake basins were relatively stable. Black Tern numbers declined in the Lake Ontario, Erie and Huron basins and declined in the Lake Michigan basin after a peak in 1996 (Figure 14b). Tree Swallow and Red-winged Blackbird numbers both increased on routes in the Lake Ontario basin (Figures 14c,d). Tree Swallow numbers declined in Lakes Erie, Huron and Michigan but were generally stable in Lake Superior. Blackbird numbers declined in the Lake Erie and Michigan basins but were relatively stable in Superior and Huron basins. Chimney Swift numbers increased in the Lake Ontario and Erie basins, with the trend in the Ontario basin a steady increase through the full five years (Figure 14e). A similar steady increase in Common Yellowthroat was observed in the Lake Ontario basin but increasing numbers were also observed for routes in the Huron and Michigan basins (Figure 14f). A B C D E F Figure 14: Annual population indices for selected bird species, summarized for those lake basins in which species were detected on at least 15 routes, 1995 through Species selected are a sample of those with significant changes in population indices observed at the Great Lakes basin-wide scale and are presented to illustrate varying spatial scales of declines or increases in numbers. 22

14 The extent to which additional years of monitoring might be expected to provide good resolution on abundance trends was assessed for marsh nesting and aerial foraging birds based on MMP data collected in 1995 through The annual trend (i.e. percent change in population index based on counts) that could be detected was calculated assuming that either 100 or 200 routes were monitored over 11 years (a 10-year interval) (Table 6). Although a standard has not yet been determined, a 3% annual trend is considered by many bird monitoring specialists as a reasonable criterion for adequate resolution of bird trends. Using this criterion and assuming at least 100 routes are surveyed for 11 years, good trend resolution is expected for 14 of the 28 species commonly recorded on MMP routes (Table 6). Included in this group are several species associated with deeper water wetland habitats (e.g. Pied-billed Grebe and Common Moorhen) and many others associated with dense wetland vegetation (e.g. Virginia Rail, Marsh Wren, Common Yellowthroat, Swamp Sparrow and Redwinged Blackbird). Monitoring on about 200 routes will be required to adequately assess trends for two species of some conservation concern, Sora and Black Tern (Table 6). Resolution on trends of a few species (bitterns, American Coot, Northern Rough-winged and Bank swallows) is not expected to meet the 3% threshold even with 200 routes. On average, about 150 routes were surveyed annually between 1995 and 1999, suggesting that the current scale of MMP monitoring is adequate for most but not all of the 28 species commonly recorded by MMP surveyors. MMP Results Table 6: Estimated percent annual change in population index (based on counts inside standard MMP station boundary) that could be detected with 100 or 200 routes surveyed over a 10-year interval (11 years of monitoring). Estimates are presented for the aerial foraging and marsh nesting bird species commonly detected on MMP routes, 1995 through Black Tern: J. Shore Detectable percent annual changes were based on Poisson regression, alpha and beta were set to 0.05 and 0.80 respectively, and estimates were adjusted for the frequency of occurrence of each species in MMP surveys, 1995 through

15 Two MMP bird survey visits are necessary MMP surveyors are asked to conduct two bird survey visits each year. Minimizing the amount of effort required of MMP volunteers is an important priority but is balanced by the need to maintain the program s ability to monitor population trends. Assessing the degree of overlap in the results from the two bird survey visits provides insight into the importance of two visits to obtaining a good sample of the bird community. Although more detailed field studies are necessary to examine the efficiency of multiple survey visits (an MMP special research project that is currently underway), some insight can be gained through the summary of MMP data presented in Table 7. For the marsh nesters and aerial foragers recorded in at least 3% of station-years, Table 7 indicates the visit with the highest count of individuals and summarizes the percent of stations with the species present during the first, second or both visits. For a given species, approximately equal counts during the two visits and a high percent of stations with the species present during both visits would suggest that surveys conducted around the average dates of either the first or second visit would likely be sufficient. Alternatively, high counts and station occupancy during only one of the two visits suggests that it is important to conduct surveys near the average date of that visit. The mix of these patterns in the MMP bird data (Table 7) supports the need for two annual visits. Counts for seven species did not differ between the first and second visit, but counts were higher on the first visit for 13 species and higher on the second visit for seven species. As expected, visits with high counts were usually the visit with a large percentage of stations with the species present. For example, counts for Pied-billed Grebe were significantly higher during the first visit and it was also recorded at more stations during the first than the second visit (71% versus 57%). Similarly, counts and frequency of detection was higher during the second visit for Black Tern and Common Yellowthroat. However, patterns for a few common species such as Marsh Wren, Swamp Sparrow and Red-winged Blackbird suggest that counts may differ between visits but that station occupancy is high during both visits. Many of the patterns presented in Table 7 are related to species breeding biology and how seasonal changes in their behaviours influence how detectable they are by MMP surveyors. During the 1995 to 1999 period, the average dates for the first and second MMP visits were June 5 and June 26, respectively. Many species are quite vocal earlier in the season when they are establishing territories and attracting mates, and tend to be more secretive later in the breeding season. Counts and station occupancy of normally elusive species such as American Bittern, Virginia Rail, Pied-billed Grebe, Sora, and American Coot tended to be higher during the first visit. Marsh Wren and Common Yellowthroat patrol their territories throughout the early summer and, as summarized in Table 7, tend to be visible and recorded during both first and second survey visits. Higher counts and station occupancy during the second visit for species such as the Black Tern may have been related to the greater activity of parents in foraging for themselves and their growing nestlings. Table 7: A comparison of mean counts and percent occurrences of marsh nesting and aerial foraging birds that were observed in 3% or more of station-years. The high average count indicates if there was a significantly higher mean count during the first visit or the second visit; equal signs denote those species for which counts were similar between the two visits. The last three columns indicate the percent occurrence of each species during the first, second and both visits. Differences between mean counts were determined with sign rank tests. 24