SPRING STAGING WATERFOWL ON THE NAKNEK RIVER, ALASKA PENINSULA, ALASKA, MARCH-MAY 2006

Transcription

1 SPRING STAGING WATERFOWL ON THE NAKNEK RIVER, ALASKA PENINSULA, ALASKA, MARCH-MAY 2006 Susan E. Savage Wildlife Biologist and William Murray Waterfowl Intern Key Words: Alaska Peninsula, American wigeon, black scoter, Bristol Bay, cackling goose, Canada goose, common goldeneye, common merganser, disturbance, Eurasian wigeon, greater scaup, greater white-fronted goose, green-winged teal, Kvichak Bay, long-tailed duck, mallard, migration, Naknek River, northern pintail, northern shoveler, red-breasted merganser, staging, trumpeter swan, tundra swan, waterfowl U.S. Fish and Wildlife Service Alaska Peninsula/Becharof National Wildlife Refuge King Salmon, AK January 2007

2 TABLE OF CONTENTS ABSTRACT... 1 INTRODUCTION... 1 STUDY AREA... 2 METHODS... 2 Weather... 2 Ground Surveys... 2 Historical Analysis... 3 Disturbance... 3 RESULTS... 4 Weather... 4 Effort... 4 Abundance, Timing of Staging and Composition... 4 Upper Route... 5 Lower Route... 6 Species Accounts... 8 Upper Route... 8 Lower Route Historic Analysis Disturbance DISCUSSION Disturbance Recommendations ACKNOWLEDGMENTS LITERATURE CITED ii

3 LIST OF TABLES, FIGURES, AND APPENDICES TABLES Table 1. Summary of spring weather conditions in King Salmon, Alaska Peninsula, Alaska, with historical averages...21 Table 2. Date of first observation of principal species on the upper route, Naknek River, Alaska Peninsula, Alaska, Table 3. Date of peak abundance of principal species on the upper route, Naknek River, Alaska Peninsula, Alaska, Table 4. Peak abundance counts of principal species on the upper route, Naknek River, Alaska Peninsula, Alaska, Table 5. Date of first observation, date of peak abundance, peak abundance counts of non-principal species on the upper route, Naknek River, Alaska Peninsula, Alaska, Table 6. Date of first observation of principal species on the lower route, Naknek River, Alaska Peninsula, Alaska, Table 7. Date of peak abundance of principal species on the lower route, Naknek River, Alaska Peninsula, Alaska, Table 8. Peak abundance counts of principal species on the lower route, Naknek River, Alaska Peninsula, Alaska, Table 9. Date of first observation, date of peak abundance, peak abundance counts for non-principal species on the lower route, Naknek River, Alaska Peninsula, Alaska, Table 10. Summary of disturbance type frequencies during Naknek Waterfowl Ground and Disturbance Surveys, Alaska Peninsula, Alaska, Table 11. Summary of all disturbance categories and effect on waterfowl tribes present, Naknek Waterfowl Disturbance Survey, Naknek River, Alaska Peninsula, Alaska, iii

4 Table 12. Disturbance rates, flushing rates (disturbance or disturbances causing flushing per hour), and flushing rates per disturbance by variable, Naknek Waterfowl Disturbance Survey, Alaska Peninsula, Alaska, Table 13. Potential waterfowl disturbances per hour of observation by year, type of survey and major disturbance category, Naknek Waterfowl Disturbance Survey, Alaska Peninsula, Alaska, FIGURES Figure 1. Location of the Naknek River Spring Waterfowl Survey on the Alaska Peninsula, Alaska...32 Figure 2. Map of the Naknek River Spring Waterfowl Lower Route Survey Points, Alaska Peninsula, Alaska...33 Figure 3. Map of the Naknek River Spring Waterfowl Upper Route Survey Points, Alaska Peninsula, Alaska...34 Figure 4. Total waterfowl abundance on the upper route, Naknek River, Alaska Peninsula, Alaska, Figure 5. Total waterfowl abundance on the lower route, Naknek River, Alaska Peninsula, Alaska, Figure 6. Total waterfowl abundance on the upper route by tribe, Naknek River, Alaska Peninsula, Alaska, Figure 7. Total waterfowl abundance on the lower route by tribe, Naknek River, Alaska Peninsula, Alaska, Figure 8. Waterfowl abundance for all tribes on the upper route and for sea ducks on the lower route from , Naknek River, Alaska Peninsula, Alaska...37 iv

5 APPENDICES Appendix I. Waterfowl species recorded over 15 years including relative abundance in each year, Naknek River, Alaska Peninsula, Alaska, Appendix II. Survey dates within interval-year slots that were used in the historic analysis, Naknek River, Alaska Peninsula, Alaska, Appendix III. Cell totals represent individual birds counted for one survey summed by tribe, interval and year, Naknek River, Alaska Peninsula, Alaska, v

6 ABSTRACT Alaska Peninsula/Becharof National Wildlife Refuge staff conducted a survey of spring staging waterfowl on the Naknek River in the Bristol Bay drainage, Alaska Peninsula, Alaska, from 14 March 16 May, Standardized ground surveys have been conducted on the Naknek River since 1992 (surveys were initiated in 1991), and historical information on phenology, species composition, and abundance is presented for each survey year. In 2006, we observed 23 species, including Brant (Branta bernicla) for the first time since Brant and ring-necked duck (Aythya collaris) were only seen during disturbance surveys. We analyzed the upper and lower route data separately for first arrival date, peak count, and peak date summaries. Fifteen-year averages of these data were calculated for 13 principal species on the upper route and eight on the lower route. Peak counts were high, and arrival and peak counts were late for many waterfowl species this year. We re-sampled survey data from all years to obtain waterfowl counts with a uniform effort and then made between year comparisons. Following a standard protocol initiated in 2003, disturbance data were collected to determine disturbance rates and effects of disturbance on staging waterfowl. INTRODUCTION The Naknek River (Naknek), part of the Bristol Bay drainage, represents one of the more important spring waterfowl staging areas on the Alaska Peninsula due to the large numbers of ducks, geese and swans that congregate here during migration. In spring, the Naknek is one of the first large water bodies to be ice-free on the northern Alaska Peninsula (Wilk 1987), and therefore represents a critical stop for migrating birds headed to northern breeding grounds. The river also supports large numbers of wintering ducks (especially common mergansers [Mergus merganser] and goldeneyes [Bucephala spp.]) during the early spring. Later in the spring, the Naknek also supports numerous species that return each year to breed on the Alaska Peninsula. Waterfowl ground surveys have been conducted by Alaska Peninsula/Becharof National Wildlife Refuge staff since This report presents data from 1992 (when survey methodology was standardized) through The purpose of the ground surveys is threefold: 1) to establish a historical record of species composition, abundance, and phenology of the spring waterfowl migration; 2) to monitor annual variations, especially in abundances of common species; and 3) to relate annual variations to factors such as human disturbance along the river and migration phenology. Tracking long-term changes in migration patterns could be highly informative examined across many staging sites over many years. This comprehensive picture may help detect trends across flyways. On the Naknek specifically, annual variations may also be revealing. Human disturbance on the Naknek is expected to increase from current levels due to the inception of a spring waterfowl hunt in 2003, a growing sport-fishing and wildlife watching tourism industry, and new construction (including plans for a bridge from Naknek to South Naknek). In addition, air traffic is fairly heavy due to the proximity of the airport with Paradise Point (one survey point) where waterfowl concentrate in large numbers. Other threats to waterfowl habitat include recent oil and gas leasing and potential development on the Bristol Bay lowlands, 1

7 possible oil leasing offshore in Bristol Bay, and a heavy metal mine in the upper Kvichak and Nushugak drainages. Disturbance data was initially collected in , but survey methodology was not standardized until Long-term data from all parts of this survey are critical to understanding changes in habitat availability to staging waterfowl as well as providing information for decisions on future development in the region. STUDY AREA The Naknek River is located toward the northern end of the Alaska Peninsula and connects the communities of King Salmon, Naknek, and South Naknek (Figure 1-3). It lies north of Becharof National Wildlife Refuge and west of Katmai National Park and Preserve. The river drains Naknek Lake, which lies within Katmai National Park, and flows westward for 51.2 km before emptying into the northern end of Bristol Bay. The lower river, from the mouth to Flat Nose Henry Road, is tidally influenced with its banks characterized mainly by upland shrubs and wet and dry tundra. The upper river (upstream of Flat Nose Henry Road) is tidally influenced to Rapids Camp, 35.2 km from the mouth, and flows through mixed boreal forest, shrubland, and wet and dry tundra. The river cuts through glacial deposits and significant earthen cliffs are exposed, especially below Paul s Creek, at Paradise Point and the Big Creek Overlook. On both the upper and lower river, extensive mud flats are exposed during low tides. Other major rivers near the Naknek, which also drain into Bristol Bay, are the Kvichak to the north and the Egegik and Ugashik to the south. METHODS Weather Weather data including average monthly temperature and monthly precipitation for March, April and May were obtained from the National Weather Service Office in King Salmon. Historical data were obtained from the Western Regional Climate Center (National Oceanographic and Atmospheric Administration) web site. This year s monthly averages were compared against 50 year averages for the area. Ice-out dates were collected at the USFWS dock in King Salmon by the first author. Ground Surveys The general survey protocol, including detailed point descriptions, is found in the Refuge s Wildlife Inventory Plan: Naknek Spring Waterfowl Survey (USFWS 2004). Methods are those used in 2004 (Schuster 2004) and generally through the history of the survey; as in the past, the surveys were divided into an upper and lower route (Figures 2 and 3). Data from the newer survey points, Oxbow (survey point [SP] 16B) and Flat Nose Henry Rd. (SP 9B), were not used in the reported sums, averages, peaks or first observation date. However, any unusual sightings made at these locations are reported. In 2006, Naknek City Dock (SP 4) counts continued from the office level due to safety concerns when barges begin unloading there in May. Also in 2006, Oxbow was inaccessible due to deep snow until 6 April and Wolverine (SP 8) was also inaccessible due to deep snow until 12 April. It appeared that sections of river were still frozen on the survey previous to the accessible date. 2

8 For data analysis, upper and lower route data were separated. Any birds unidentified to tribe (approximately 216 birds on the upper route or 0.3% of total counts) were excluded from sums when calculating descriptive percentages. For each route, principal species were designated (13 for the upper route, 8 for the lower route) as a core set of data to be included in descriptive statistic summaries. Descriptive statistic summaries included average peak abundance and standard deviation, as well as the average arrival date and average peak date, both with standard deviation. Averages span the 15-year period from during which the protocol has remained standardized. Historical Analysis Year to year comparison of survey totals has been problematic because of varying effort between years. To correct for this, we documented all survey dates in a worksheet (separate for Upper and Lower routes) and divided it into multi-day intervals. Our goal was to find the smallest multi-day interval with a survey in each year-interval slot. This was an eight day interval for the upper route and a 15 day interval for the lower route. A survey s data might be moved by one day from one interval to the next to complete a missing datum if necessary. When multiple surveys occurred in an interval, one was selected randomly to represent that interval. A preliminary examination of these data was conducted to look for multi-year patterns of arrival and abundance. We ran correlations of year total counts within and between tribes. We also ranked each tribe for each year by earliness of arrival and by earliness of peak. Spearman correlations (r s ) were run on ranked values. We also ran correlations of tribe totals and earliness ranks against the sum of March and April average temperature and the sum of March and April total precipitation. Disturbance We followed the protocol developed in 2003 and used since. We attempted to schedule three surveys per week (one randomly chosen weekday, Saturday and Sunday); however ground surveys and weekend disturbance surveys were prioritized. All 2006 data were entered into the ACCESS database created in As in later 2005 surveys, upon arrival at each disturbance survey, we began making estimates of size and make-up of discrete groups of birds visible from the viewing site as quickly as possible; we later filled in details if no disturbance had yet occurred. A disturbance event was defined as any external source that caused a change in the behavior of the waterfowl or any source perceived by the observer to potentially cause a change in the behavior of waterfowl. Types of disturbance events were categorized and analyzed for frequency (disturbance events per hour). Number of birds disturbed by tribe and level of effect were also summed. Calculating number of disturbances per unit of time is problematic for ground surveys, so rate analysis was conducted only for disturbance-only surveys. Rates of disturbance events, events where birds flushed (categories 3 5), and flushes/disturbance were compared by location (Paradise Point [SP 14] vs. Rapids Camp [SP 16]), month (April vs. May), weekday vs. weekend day using a two-tailed Student t-test (significance 3

9 level of 90%). Analysis of variance was used to test for differences between time of day categories. Potential observer caused disturbance has not been recorded uniformly across all years. In 2003 and 2004, observer approaches to a survey point (on foot or by truck) were only recorded as a disturbance event when birds were close to shore and/or observer was clearly visible to waterfowl. In 2005, observer approaches were entered even if no obvious disturbance occurred. In this way, baseline data (species count prior to potential disturbances) for the survey were entered for every survey, even if no disturbance occurred. In 2006, baseline data was entered for each survey. However, observer approaches were recorded as disturbance sources only when birds were affected by the observer (category 2 5), and only during disturbance (not ground) surveys. RESULTS Weather Initially the river opened on 19 March, but then re-froze on 22 March. The second and final ice-out date on 29 March was one day later than the 15-year average. March and April were cooler than average (by 5 degrees and by 2 degrees, respectively) while May was slightly warmer (almost 2 degrees, Table 1). March and May had slightly less than average precipitation while precipitation in April was more than double the average (Table 1). Effort Surveyors conducted ground-based surveys from 14 March to 16 May Thirteen upper route and 11 lower route surveys were included in the waterfowl summary for a total of 24 surveys. We recorded the time spent at each survey stop. We recorded 26 hours 53 minutes (13 surveys) of observation time on the upper route including Oxbow SP and 13 hours 45 minutes (11 surveys) of observation time on the lower route including Flat Nose Henry Road SP. Approximately 82 hours were spent in total survey time (including driving between survey points; 51 hours on upper and 31.3 hours on lower) which did not include transit time to the first survey point, transit time back from the last survey point, or time spent on several incomplete surveys (stopped due to weather conditions). Abundance, Timing of Staging and Composition Observers noted 23 waterfowl species during surveys in 2006: 16 on the lower route, 19 on the upper route (Appendix I), and two species noted only during disturbance (brant and ring-necked duck). Although the ice at King Salmon moved out as in average years, many stops below remained frozen for over a week. This undoubtedly affected abundance patterns on the lower route. On the upper route all principal species arrived later than the 15-year average (Table 2). Except for Eurasian and American wigeon (Anas penelope and Anas americana) and northern pintail (Anas acuta), other principal species peaked later than average on the upper route (Table 3). For the lower route the arrival dates of principal species varied, but most also peaked later than average (Tables 6 and 7). Common mergansers and common goldeneyes (Bucephala clangula) were 4

10 present at the outset of the survey on upper route, but goldeneyes were not observed on lower route until 12 April. These species are not given an average arrival date because the average date would arbitrarily represent the first survey date (Table 2 and 6). The frozen state of the river and moving ice flows forced common mergansers to move between the upper and lower routes in search of open water. Common goldeneyes, however remained concentrated in the upper river this year. Sea ducks were by far the most plentiful group on the lower route, augmented by divers in late April and May. Dabblers, swans, and geese were concentrated on the upper route. The total waterfowl peak count (11,550 birds) occurred on 28 April on the upper route with a previous peak (of 11,375 birds) on 24 April (Figure 4). On the lower route the peak occurred on 12 April at 3,225 birds (Figure 5). Peak abundances of all principal species on the upper route were notably higher than average (by at least one third) in 2006 except for both species of wigeon and common goldeneye (Table 4). Peak abundances of four of eight principal species on the lower route were also notably higher than average this year (Table 8). Information on first observation date, peak date and peak count for non-principal species is also reported (Tables 5 and 9). Upper Route Three shallow lagoons along the upper Naknek observable from Paradise Point, Big Creek Overlook (SP 15), and Rapids Camp - have shown high waterfowl concentrations in all years of the survey and represent critical stop-over habitat for many species, especially tundra swans (Cygnus columbianus) and northern pintails. The upper route in general supports the largest concentrations of waterfowl on the Naknek River. Geese (Anserini) This tribe is predominantly represented by greater white-fronted geese (Anser albifrons) along the Naknek. The majority of the migrating birds continue further north to breeding grounds on the Yukon-Kuskokwim Delta (YKD breeding population), although small numbers stay to breed in the Bristol Bay lowlands of the Alaska Peninsula (BBL breeding population, Ely and Takekawa 1996). Both emperor geese (Chen canagica) and brant (Branta bernicla) are observed in some years in small numbers; brant were seen in 2006 but only during a disturbance survey. This year Canada and cackling geese (Branta canadensis and B. hutchinsii) numbers (summed together) were far above historical numbers, correlating closely with the arrival date, peak abundance date, and above average numbers of white-fronts (Table 2-4). Peak abundance count for geese was 1,745 birds on 24 April (15% of total waterfowl for that day). Geese represented 9.5% of total waterfowl counted in 2006 on the upper route. Swans (Cygnini) This tribe is represented by two species, tundra swans and trumpeter swans (Cygnus buccinator). Tundra swans are the more numerous species and are primarily found on the upper route. Because of the difficulty of identifying species at a distance, counts of 5

11 tundra swan include both species of swan 1. Trumpeter swans were seen for the first time on the Naknek in Tundra swans are present in relatively large numbers on the river throughout the survey and represent a portion of the local breeding population. Counts include swans visible in ponds across the river. Peak abundance occurred on 28 April (2,679 birds), and represented 23% of total waterfowl on that date. Swans represented 19% of total waterfowl counted in 2006 on the upper route. Dabbling ducks (Anatini) Dabblers are the most abundant tribe found on the upper route and on the Naknek, and represented 43.5% of upper route waterfowl counted in Peak counts of 6,100, 6,669 and 5,538 occurred on 24 April, 28 April, and 9 May; the vast majority of these were northern pintails. Smaller numbers of breeding dabblers remain on the river past the end of waterfowl surveys. Dabblers numbered 373 birds (29% of daily count) on the last day of the survey, 16 May (Fig. 6); northern shoveler (Anas clypeata) made up 32% of dabblers that day. Diving ducks (Aythyini) Historically, diving ducks arrive and peak later (with much lower peak numbers) than the other waterfowl tribes. Since other tribes were late this year, diver peaks were more similar to other tribes. They contributed 9% of waterfowl counts on their peak day 9 May, and represented 4% of upper route waterfowl counts. On the upper route divers are mainly represented by greater scaup (Aythya marila) observed at Grassy Point (SP 13). However, on their peak date, numbers were also high at Paradise Point and Big Creek Overlook. Sea ducks (Mergini) Sea ducks are represented mainly by common mergansers and common goldeneyes. Both species winter in the Bristol Bay region and were present on the first survey on 14 March with a combined count of 183. The lack of open water on the first survey accounted for low numbers; common mergansers were frequently counted on the lower route early in the survey. Sea duck numbers gradually rose as water opened from Lake Camp (SP 17) to King Salmon. The peak for sea ducks occurred 6 April with 2,797 birds and 90% of the count that day. Most sea ducks, except red-breasted mergansers (Mergus serrator), left the upper route for breeding grounds by late April as dabblers, swans and geese numbers peaked. Sea ducks accounted for 24% of ducks counted on the upper route. Lower Route During the months of this survey, the lower route supports fewer numbers of waterfowl than the upper route; 3,226 birds were counted on the peak (12 April) versus more than 11,000 for the upper route (24 and 28 April). However, because of the frozen conditions of the upper river early in the survey, the lower river provided critical resources for common mergansers. Local weather conditions have more influence over survey totals 1 On April 11 trumpeter swans were distinguishable from tundra swans and were recorded separately at Grassy Point which allows closer views than other survey points. 6

12 on the lower route; wind and fog can reduce visibility especially at the State Road (SP 3), Sewage Lagoon Overlook (SP 2), and Kvichak Bay (SP 1). Geese Geese were only observed on 26 April at two lower route survey points: Flat Nose Henry Road and Wolverine. Excluding totals from Flat Nose Henry Road, only two Canada geese were observed. Geese represented less than 1% of total waterfowl counted on the lower route. Swans Swans were observed at the upper two survey points of the lower route; only tundra swans were observed. Excluding swans observed at Flat Nose Henry, swans were observed only on 15 May and the total lower route count for the tundra swans was 12. Swans represented less than 1% of total waterfowl counted on the lower route. It is probable that the low numbers indicated a lack of preferred food sources. Dabbling ducks Dabbler numbers are strongest in the upper part of the lower route. Dabblers represented only three percent of the total waterfowl counted on the lower route. The tribe reached its peak abundance of 188 on 26 April; this date was between the waterfowl peak dates on upper route. Dabblers were represented by six species with northern pintail making up about two-thirds of the numbers. On the last day of the survey, American wigeon contributed to almost half the dabbler numbers on lower route. Diving ducks Divers represented 15% of total waterfowl counted on the lower route, and reached their peak abundance of 529 on 10 May. Diver numbers remained high to the last survey day, although because of the unusually high number of red-breasted mergansers, they did not dominate the totals as they did in As on the upper route, greater scaup represent 99% of the diving duck total with a small contribution from canvasbacks (Aythya valisineria). They were most abundant at the Kvichak Bay site except on the last survey day where almost half were at Leader Creek (SP 5). Sea ducks Sea ducks were by far the most plentiful tribe on the lower route, accounting for 82% of total waterfowl counted. On 12 April, they reached a peak abundance of 3,210, the majority of which were common mergansers. This was close to their peak date for the upper route also (6 April). This high count of common mergansers was second only to those counted in Long-tailed ducks (Clangula hyemalis) were not as numerous as in Sea duck species peak date on the lower route tend to be more variable than other species on either route (Table 3 versus Table 7, note standard deviations). Because of unusually high numbers of red-breasted mergansers on the last survey day, sea ducks represented 49% of the total lower route counts. Sea duck species diversity also rivaled dabbler diversity with six species on 5 May. 7

13 Whereas swans, geese and other ducks are few at the beginning of the season, peak and then decline, sea ducks as a tribe have a variable pattern. Some species are present at the beginning of the season and decline, while other species arrive, peak and decline, and other species may irregularly flood survey totals during any part of the season (i.e., see king eider [Somateria spectabilis] count in 1994 [MacGowan 1994 ] and long-tailed duck counts in 2005). This year because of movements of common mergansers between the upper and lower route, sea ducks actually peaked mid-season on the lower route. Species Accounts Upper Route Principal species are designated by blue text. Annual and average statistics for first and peak arrival date and peak count are presented for principal species (Tables 2-4) with only 2006 data presented for non-principal species (Table 5). Relative abundance for all years is presented in Appendix I. In 2006, many species, especially dabblers, geese and swans, arrived and peaked later than average and many peak counts were high. Greater white-fronted goose, the most common species of goose on the Naknek River and upper route, deviated from its average arrival and staging pattern this year. White-fronts were first noted on 14 April, six days behind the average arrival date, and peaked on 24 April, three days later than average. The peak count of 1,537 was well above the 15-year average of 956, but below the 2005 peak of 2,425 birds. White-fronts usually stage on the Naknek for about three weeks, but this year were abundant for less than two. Canada geese and cackling geese are totaled together for this survey. Formerly known as Canada goose, several smaller subspecies of this diverse group were re-named cackling goose by the American Ornithologists Union in 2004 (Banks et al. 2004), while the larger subspecies remained as Canada goose. Since observers could not always differentiate between these species, accurate numbers for the two are unavailable. In addition, historical data is not available for what was formally considered separate subspecies. We recorded the first Canada goose on 14 April; the first cackling geese were differentiated on 24 April. Numbers peaked on 28 April with 290 birds; no Canada or cackling geese were seen after this peak date. This peak is more than two times the 15-year average of 101 birds and is greater than last year s peak and record high of 221. Tundra swans are a common species on the Naknek. This year the peak count exceeded all other years (2,679) and was significantly above the 15-year average (1,415). This species showed less variation from average in arrival (25 March) and peak dates (28 April) than other principal species this year. Tundra swans remain on the upper Naknek in higher numbers than many species toward the end of surveys; observers counted 299 swans on the last survey. Local observers and the survey team observed several collared swans in 2004, 2005, and Three individual collared swans were observed this year, beginning on 23 April. Observed residence times ranged from 1-43 days (Richard Russell, p.c.). Two of the collared birds sighted this year were sighted last year, but none were resighted from In 2004 we learned that some of these birds were banded incidentally in Washington 8

14 State s Whatcom County as part of a local project to determine the source of lead poisoning in trumpeter swans. Recently M. Smith (p.c.), one of the investigators on the Washington project, confirmed that eight of ten collared birds sighted from 2004 to 2006 were banded during this project. We learned the other two birds were banded at the Stillwater National Wildlife Refuge in Fallon, Nevada in 2001 (B. Henry, p.c.). Trumpeter swans were observed on the waterfowl survey for the first time in 2004, and there is no evidence that they have been seen before on the Naknek during spring migration (Schuster 2004). Trumpeters are classified as an uncommon to rare migrants in western southcoastal Alaska (Bellrose 1980, Kessel and Gibson 1978). In 2006, four trumpeters were positively identified on 14 April. On other surveys, trumpeters were not specifically noted. Local observers following trumpeter swan arrival first noted them at Lake Camp on 21 February. Gadwall (Anas strepera) are uncommon spring migrants and arrive relatively late in the spring. Three individuals were sighted 9 May. Eurasian wigeon occur regularly in small numbers on the survey, and most sightings appeared to be unpaired males. The difficulty of distinguishing female Eurasian and American wigeon has been previously noted (Meixell and Savage 2004). This year, the second author and primary observer assumed that there was a hen for each drake Eurasian wigeon. He generally observed drake and hen numbers to be similar. In cases where there were more hens than drakes, he weighted the hens to the species of wigeon that was more common. The Eurasian wigeon s first arrival date (24 April) was almost two weeks past the 15-year average, while the peak date came quickly after arrival and was three days before (28 April) the 15-year average peak. Eurasian wigeon were one of the few species whose peak number did not vary from the 15-year average (30 birds counted in 2006 versus 15-year average of 31). American wigeon arrived one and a half weeks later than their 15-year average of 10 April but peaked six days ahead of the 4 May average. They are moderately abundant compared to other dabblers. This year s peak count of 218 was about 10% above the 15- year average. Mallards (Anas platyrhynchos) were not present early in the survey as in They first were noted on 6 April, a week after their 15-year average, and peaked on 28 April, 11 days after their 15-year average peak. They are the second most abundant dabbler on the Naknek River. This year s peak of 895 birds was more than twice the 15-year average peak and was greater than the previous high noted in 2005 of 828 birds. Northern shovelers generally arrive and peak relatively late in the year compared to other dabblers. In 2006, they arrived on 24 April, two days past their 15-year average and peaked on 9 May, one day after their 15-year average peak. Of the dabblers documented, they have the lowest variance in arrival date but a high variance in peak counts (CV of 97%). Noteworthy this year was the abundance of shovelers; they peaked at 728 birds 9

15 which is over four times their 15-year average peak and exceeds the previous high of 288 noted in Northern pintails are the most abundant waterfowl species on the upper route. This year their peak count of 5,315 greatly exceeded the 15-year average peak of 3,853. It was also the third highest count over all years. Pintails arrived on 6 April, a week after their 15- year average arrival date, but peaked on 24 April, four days earlier than the 15-year average of 28 April. It is worth noting that pintail peak numbers over the 15 years of this study have been extremely variable with a range from 948 to 11,480 birds and a standard deviation of 2,623. Variance around first arrival date is the greatest of the dabbler species (standard deviation [SD] 10.6 days) and peak date is also relatively variable (SD 6.9 days), indicating that they have a less rigid timeline for migration and may time movement with environmental factors as opposed to calendar date as some geese appear to. Green-winged teal (Anas crecca) arrived on 24 April, nine days later than their 15-year average. Their peak count on 9 May was a week later than average and, at 358 birds, was high compared to the 15-year average of 194. The peak count for 2006 was second only to the 1992 count of 543. Canvasbacks were observed on three survey dates. The first observation occurred on 24 April and they peaked at 17 birds on 28 April; four birds were seen again on 12 May. Greater scaup are the only commonly occurring diver on the Naknek and were seen in comparatively high numbers in They peaked at 673 birds this year, more than twice the 15-year average of 262. Their numbers are the most variable of the principal species on the upper route, with peak counts ranging from 52 to 989 and a standard deviation equaling than the 15-year average (SD 258, coefficient of variation [CV] 99%). They were first observed on 14 April, 12 days past their 15-year average arrival and peaked on 9 May just two days past their 15-year average peak date. Bufflehead (Bucephala albeola) were observed just twice on the upper route in Six birds were observed on 28 April and four birds were observed on 12 May. Common goldeneyes are winter residents in Bristol Bay and, after common mergansers, are the most common species seen during March and early April surveys. They peaked on 21 April, 12 days past the 15-year average peak, but with a similar count (612 birds) to their 15-year average. Goldeneye were present on every survey and remained abundant through the end of April, after which they quickly declined in numbers. Barrow s goldeneyes (Bucephala islandica) are seen in small numbers on the Naknek, and were known to over-winter in Bristol Bay with common goldeneyes and mergansers (R.R., p.c.). They peaked at 7 birds on 17 March. Their numbers may be underestimated due to identification difficulties at long distances. Observers this year noted conflicting field marks: purple sheen seen on birds with large facial patches. 10

16 Common mergansers are winter residents in Bristol Bay. Common merganser numbers were relatively low on the upper route due to little open water, and increased toward the end of March. In 2006, peak abundance occurred 6 April, which is five days later than the 15-year average. Their peak date is the most variable of the principal species (SD 12.9). This year s late peak on the upper river is probably due to frozen conditions and use of the lower river by the birds. The 2006 peak count (2,303) exceeded the 15-year average peak (2,124). Red-breasted mergansers, in contrast to 2005, were relatively abundant in 2006 (peak count 265; 15-year average 182). Their arrival and peak dates and peak counts are variable compared to all principal species (SD 13.9, 12.4 and 160, respectively, CV 88% on peak count). Their arrival on 14 April was ten days later than the 15-year average in 2006 and the peak count on 9 May was almost two weeks past the 15-year average. On 13 May, during a volunteer bird count covering a similar area to both upper and lower routes, 469 red-breasted mergansers were counted. Lower Route Principal species are designated by blue text. For analysis, seven species were originally designated as principal species for the lower route. Due to the increase in the abundance of the mallard on the lower route, the species was included as a principal species in Of the eight principal species, five are sea ducks. Annual and average statistics for first and peak arrival date and peak count are presented (Tables 6-8) for principal species with only annual data presented for non-principal species (Table 9). Relative abundance for historic years is presented in Appendix I. For arrival dates there was no general pattern in However, six of eight principal species peaked late and had high peaks. Contrary to the pattern shown in 2005, the one species that peaked significantly early had a lower than average peak count. Four of six species that peaked late had significantly greater (> 50%) peak counts. Canada geese were only observed at the Flat Nose Henry Road point (which was not included in total counts) and Wolverine Rd. point on 26 April. Totals at Wolverine include just two birds. Tundra swans were observed on four days at Flat Nose Henry Road point (which was not included in total counts) and only on 15 May at Paul s Creek (SP 9A). The Paul s Creek birds totaled 12 birds. Eurasian wigeon were an infrequent visitor on the lower route and were seen only at sites up river from Martin Monsen Park (SP 6). Their peak count of four was observed on both 20 April and 15 May (dates averaged to 2 May; also the first and last days observed). American wigeon were observed in relatively small numbers and peaked at 24 birds on 15 May. They occurred at all sites upriver from the Naknek City Dock. 11

17 Mallard abundance returned to pre-2004 levels in This year they arrived on 7 April, half a week before the 15-year average and peaked on 26 April half a week later than the 15-year average. These dates were comparable to upper route dates. The peak count of 35 was below the 15-year average. Mallards were generally observed above the Savanoski crossing (SP 7), but occasionally were present all the way to Kvichak Bay. Northern shovelers remained an infrequent visitor to the lower route, in contrast to their unusual abundance this year on the upper route. They peaked at six birds on 5 May. Northern pintails occurred regularly on the lower route but in much lower numbers than on the upper route. They are most common on the upper three stops (including Flat Nose Henry Road). They were first observed on 12 April which is their 15-year average arrival date. They peaked at 126 birds on 26 April, 25% above their 15-year average, and just two days ahead of the 15-year average. Although the total count was well below the 2005 number, they ranked as the sixth most abundant species on the lower route. Green-winged teal peaked on the lower route at 15 birds on 10 May. They were first observed on 26 April and on three other survey dates. They are most common on the up river stops of the lower route. Canvasbacks were observed only on 5 May with a count of 13 birds. Greater scaup were the second-most numerous species on the lower route, with a total of 2,099 birds and 14% of total counts. They arrived on their 15-year average date of 12 April and peaked on 10 May, half a week later than the 15-year average. The peak of 529 birds was third only to last year s high count of 530 birds and the 1992 count of 796 birds. After their peak, they retained a high count through the remaining survey on 15 May. The Kvichak Bay survey point contained 81% of the total count for the greater scaup. They were the most numerous species (516 of 1,154 birds) on the last survey day. Black scoters (Melanitta nigra) were first observed at the Kvichak Bay site on 26 April, 11 days after their 15-year average. Their peak count of 129 birds occurred on 15 May, the last survey day, five days after their 15-year average. Black scoters and long-tailed duck were the latest peaking species. This year s peak was 22% above the 15-year average. Most scoters were observed at the mouth of the Naknek and in Kvichak Bay (SP 3 to 1). Long-tailed ducks are known to winter in the Bristol Bay region and have been observed in March in all years and on the first survey date in many years, so their average date of first observation is probably reflective of the survey start date. The long-tailed duck was the fourth-most numerous species on the lower route in This year they were observed on the first survey and peaked on that date, almost a month ahead of the 15-year average. The peak count of 275 birds was well below the 15-year average. During an incidental observation by the first author on 22 April at Kvichak Bay, approximately 540 birds were counted with numerous other waterfowl unidentifiable due to distance. This 12

18 species appears to be highly variable in timing of movements and staging areas. Birds may feed further offshore out of range of the survey spotting scope. Bufflehead were observed at three stops (including Flat Nose Henry Road) on 5 May for a total of 14 birds (not counting Flat Nose Henry Road birds). Common goldeneyes were accounted for less than 1% of waterfowl totals on the lower route. From 1999 to 2005 common goldeneyes accounted for 2-4% of totals. Prior to 1999 in half the years they accounted for less than 1%. This species used the upper river to a greater extent this year. They were not observed on the lower route until 12 April which was also their peak day, four days past the 15-year average. Although their season totals were low, their peak count of 101 birds was the highest ever observed for this species on the lower route. After this peak, their numbers dropped significantly and remained low through the end of the survey. The Paul s Creek site accounted for 77% of the total common goldeneye counts. Common mergansers were the most numerous species on the lower route, representing 59% of the total waterfowl counted. They were unusually abundant this year. They peaked on 12 April, two weeks past their 15-year average peak date. The peak count of 3,094 was second only to the 1992 peak count of 3,270 birds. Through the end of April, except on 4 April when they may have been upriver (high count on 6 April for upper route), they were the most common species of waterfowl on the lower route. Because the upper stops of the lower route were frozen until early April, mergansers were distributed on the lower half of the lower route during the early surveys and moved to the upper section of the lower river on later survey dates. Common mergansers did not start to decline significantly until 5 May. Red-breasted mergansers were first observed on 23 March, two weeks ahead of their 15- year average, but peaked on 10 May much behind their 15-year average peak date of 23 April. Their peak count of 863 birds is the highest ever observed, followed by the count of 822 birds in High numbers (116 and 74 respectively) of this species were observed at Flat Nose Henry Road on 5 and 10 May (not included in totals). Counts exceeding the 15-year average for lower route were noted from 5 to 15 of May. Also see the comment on red-breasted mergansers for upper route. Historic Analysis The upper route survey dates from 1992 to 2006 ranged from 26 February to 21 May (Table 2). When divided into eight intervals of eight days there were just six intervalyear slots missing data (Appendix II). A survey in the adjacent interval was used in three cases to fill these empty slots because the date was only one day more or less than the interval boundaries (Interval 1 filled, 1994; Interval 2 filled, 2000; and Interval 2 filled, 2004). After using data from a previous or next interval in these three cases, there were three missing interval-year slots (one for Interval 1 [1992] and two for Interval 8 [1993 and 2002]). Data from an incidental survey (North American Migration Count) was used 13

19 to under-estimate 2 tribe totals for Missing data in Interval 8 for 1993 could greatly affect total counts or rankings, so 1993 was excluded from correlation tests. Missing data from Interval 1 for 1992 would likely only affect sea-duck totals and it is possible that even these were unaffected by the cold and icy conditions in that year. Individual bird totals summed by tribe, interval and year were prepared separately for the upper and lower routes (Appendix III). A few general patterns emerge from these data. On the upper route, in cold years (relative temperature defined by summing March and April mean temperatures within a year) swan, dabbler and diver numbers are often high and sea ducks are slow to depart (Figure 8a, b, c). Geese have the shortest staging period especially in colder years. Correlation analysis with environmental factors compared the sum of mean monthly temperature for March and April or a sum of total precipitation for March and April against total yearly tribe numbers or a rank for arrival or peak patterns. Warmer temperatures were correlated with earlier arrival for dabblers and earlier departure for sea ducks (p < 0.05, r s = 0.77 and r s = 0.81) and negatively correlated with diver and swan total numbers (p < 0.05, r s = and r s = 0.53). Precipitation was correlated with swan numbers and weakly correlated with dabbler and diver numbers (p < 0.05, r = 0.60, p < 0.10, r = 0.45 and r = 0.47 respectively). Comparisons were also made between tribe totals and tribe arrival or departure patterns. Comparing tribe counts, swan annual totals correlated with dabbler and diver annual totals (p < 0.05, r = 0.71 and r = 0.53) respectively. Weaker correlations (p < 0.10) were observed between swan and sea duck total counts (r = 0.46) and sea duck and dabbler total counts (r = 0.47) (i.e., in years when the subject tribe was abundant, so was the other tribe). When rankings of each tribe s total counts are compared to arrival date patterns (or departure date for sea ducks), the only significant correlation (p < 0.05) was within sea ducks (r s = 0.54) (i.e., in years with low sea duck totals, birds departed earlier). Comparing arrival ranking between tribes, dabbler and diver arrival ranks were weakly correlated (p < 0.10, r s = 0.46). Dabbler and diver and diver and goose peak rankings were also correlated (p < 0.05, r s = 0.77 and r s = 0.64 respectively), while dabbler and goose peak ranks were weakly correlated ((p < 0.10, r s = 0.51). These last correlations indicate that these tribes arrived and/or peaked at similar times on the upper river. The lower route survey dates from 1992 to 2006 were inclusive from 1 March to 29 May (Table 6). In several years, lower route surveys were conducted less frequently so longer intervals had to be selected to have relatively complete coverage of interval-year slots. Five 15-day intervals were selected. Data was missing for four years for interval five (1993, 1995, 2000, and 2002). A survey from the adjacent interval was moved for 1995 and 2000 (Appendix II). The other two years were eliminated from the correlation tests. Swan, goose and dabbler numbers are so low on the lower route that correlations were not run for these tribes. In addition, fewer intervals resulted in a decreased ability to find relationships between data. We noted that sea duck numbers may be driven by temperature to a greater extent on the lower route (Figure 8c). The only significant 2 The survey was conducted by teams which were not strictly divided by the upper and lower route designation of this survey. Data was not used from one team which collected information on both upper and lower route stops. 14

20 correlation found was between sea duck and diver annual total counts (p < 0.05, r = 0.67). Weak relationships were found between diver numbers and precipitation, sea duck numbers and precipitation, and diver numbers and diver departure patterns (p < 0.10, r = 0.48, r = 0.51 and r s = 0.49 respectively). Disturbance Disturbance surveys began on 10 April at Paradise Point and 19 April at Rapids Camp after the final ice-out was complete. The last survey was on 23 May. Twenty-three hours of observations were collected during disturbance-only surveys. Approximate effort during the ground surveys would include 40 hours 38 minutes as estimated above for total ground observation time. A disturbance occurred on each disturbance-only survey and on four of 185 survey points (not counting stops where the river was frozen) during 24 ground surveys (including Flat Nose Henry Road and Oxbow). As stated in the 2005 report, these data probably underestimate the number of disturbances occurring during regular ground surveys. The following summary presents combined data from both survey types. A total of 116 disturbance events were documented: 80 human-caused, 25 natural, and 11 unknown. Observer related disturbances were not recorded during early surveys in The largest percentage of disturbances involved hunting; there were 40 hunting events (34.5%) most of which included hearing gunfire without visual confirmation. Aircraft accounted for another 21.6 % of the total disturbance events. Predatory birds (bald eagles, common ravens and gulls) also accounted for 21.6% of total disturbance events. Vehicles and other (including unknown) accounted for 11% and 10% respectively (Table 10). To illustrate how many birds per tribe responded to types of disturbance events at any level of effect (category levels 2-5), the number of birds present before each disturbance were compared to the number of birds responding (Table 11). Of all the waterfowl available on the river to be disturbed (approximately 113,800 birds over all surveys), less than three percent responded during the 116 events we observed. During surveys when disturbance was recorded, tribes were present in the following abundance: dabblers (51%), swans (22%), geese (12%), unknown (9%), sea ducks (6%) and divers (1%). Because of their high numbers, by raw number, dabblers were disturbed more than all other tribes combined. However, when examined by percent of birds present, the tribe most affected by disturbance this year was sea ducks (5.6% of sea ducks present were disturbed) followed by geese (3.7% of geese present were disturbed). Seventeen events resulted in the disturbance of 50 or more birds. These included: two over flights of medium aircraft, three adult bald eagle (Haliaeetus leucocephalus) flushes, one gull flush, the hunting from blind event, the conspicuous person event, two shots fired events, three events of unknown cause, one ATV event, and three boating events. These disturbance events causing 50 or more birds to respond also resulted in higher level of disturbance (categories 3 5). The following results address data from disturbance-only surveys. Unless stated to be statistically significant, the following observations are not so. Rates of disturbance events, rates of disturbance events that caused birds to flush (effect categories 3-5; both calculated per hour), and flushes per disturbance are reported along with probability (p) 15

21 values for t-tests, F value for ANOVA (Table 12). Disturbance events per hour were higher in April than in May (p = 0.04), however more birds flushed per each disturbance event in May (P = 0.02). The two survey locations did not differ in disturbance rates or responses. Birds flushed more frequently per disturbance during weekend disturbance events (p = 0.04). No surveys were conducted during the morning time slot ( ). No correlation was found between time and any calculated disturbance rate. When grouped by time of day categories (midday = , afternoon = , evening = ), Analysis of Variance detected no differences between categories for any disturbance rate calculation. DISCUSSION Although the final ice-out date of 29 March is similar to the 15-year average, 2006 was anything but an average year. Conversely, in 2005 a later than average ice out date did not reflect the warm spring. In 2006, the average temperatures for March and April show the area experienced a cold and a late spring. A seven-year pattern has emerged over the 16 years of this survey which is evident in the general weather pattern and in some species arrival dates, peak dates and abundance. This cold pattern was clearly evident in 1992, 1999 (when ice remained in the Rapids Camp lagoon until late May), and this year. In contrast to 2005, many species of waterfowl arrived late and peaked late. However similar to 2005, many peaks were higher than average and there was an abundance of waterfowl (especially geese [but for a shorter window], swans, most dabblers, greater scaup, and mergansers) on the river. It is likely that species were late because of the frozen conditions of the Naknek and of late phenology further south delaying their migration. Different species may be affected to different degrees depending on their habitat use patterns. It also appears that some species stayed on the river longer or used the Naknek as an alternative staging area to other waterways that may have been frozen. Bethel (near the Yukon-Kuskoquim Delta and breeding grounds for many waterfowl species) weather demonstrated the same pattern found in King Salmon with March and April temperatures well below average, and April above average in precipitation. Kodiak temperatures, but not precipitation showed a similar pattern. This is the third consecutive year we have observed collared tundra swans arriving from the state of Washington. Ten individual birds 3 have been observed over three years with four observed in 2004, eight in 2005, and three this year. In 2006 we observed two repeat birds from 2005, and in 2005 there were three repeats from 2004, but no individual has been observed in all three years. In 2005 we observed a mated pair both of which were collared; only one of those birds was observed in Swans were banded and collared statewide as part of the Avian Influenza monitoring effort in the summer of 2006 including 64 birds at Caribou River south of Port Moller and 95 birds on the northern Alaska Peninsula (Savage and Sowl, in prep.). Hopefully this will lead to additional sightings of marked swans in the spring of This is also the third consecutive year we have observed trumpeter swans on the Naknek River. Trumpeters were the first swan species to arrive on the river. Juvenile trumpeters were observed again in Counting F13 (red), originally identified as a trumpeter swan, as a tundra swan. 16

22 Disturbance Historical information pertaining to spring disturbance of waterfowl on the Naknek that predates the disturbance survey can be found in Schuster (2004). When comparing rates of potential disturbance events between years (Table 13) what is most striking is the between year variation. This year the total rate of potential disturbances and the occurrence of bald eagle, medium and small aircraft, and hunting events was greater than in most other years. This variation in part is caused by the low number of surveys and the random occurrence of disturbance events. This year the measurement of observer disturbances differed and there were no early morning surveys also adding to variation in detecting potential disturbances. One should note that disturbances during ground surveys were likely under reported again in 2006 (see Lapinski and Williamson 2005). Total percentage of birds responding and calculated disturbance rates provide two different measures of disturbances and bird responses to disturbance. Although the total percentage of birds responding (categories 2 5) to disturbances dropped from 2005 to 2006 (7% to 2.6%), the rate of potential disturbance events observed, the disturbances that caused flushing (disturbance categories 3 5), and the rate of flushes/disturbance were higher in 2006 and more similar to those observed in The measures flushing and flushes/disturbance are triggered even when small numbers of the total birds present respond to the disturbance, so these measures can be higher even when numbers of total birds disturbed are lower. In addition, other biological or behavioral factors may be at work. Large numbers of birds as observed in 2006 may offer safety in numbers so fewer birds react to some disturbance events. In addition, more disturbances may cause some birds to habituate to certain types of disturbance. Disturbance data collected for this survey may prove vital to future waterfowl conservation on the Naknek. While other rates of disturbance events vary between years, these data (and Anecdotal information - see Appendix 2, Schuster 2004) show an increase of hunting events on the Naknek since the beginning of the spring waterfowl season. Hunting events increased from 0 in 2003 (57 observation hours), 11 in 2004 (67 observation hours), 17 in 2005 (88 observation hours), to 40 in 2006 (67 observation hours). Considering the fact that there were no early morning observations in 2006 when hunting is often most common, the 2006 data possibly underestimates the trend. This year one visual observation during surveys of hunting was made. Because gunshots can be heard from varying distances depending on weather conditions, disturbance to surveyed birds, which are probably reacting to a much lesser degree than those at the source of the disturbance, is not a good measure of direct affect. The Naknek River provides a critical stop-over area for migrating birds to rest and refuel before moving to breeding areas north and south of the river. Over the 16 years of this study, considerable variation has been observed in timing, species presence, and species abundance. However, patterns dependent on weather are emerging. The complexities of these patterns and how they affect large scale populations are poorly understood. These data are available to any scientist who may provide further insight into the patterns they represent. 17

23 Recommendations Due to budget shortfalls and the difficulty in recruiting an intern for this early survey, it is likely that the survey will not be conducted annually by Refuge staff in the future. This does not preclude future work on this topic. The following, some of which were proposed in 2005, are valid suggestions for related work: 1) Following the historical analysis conducted in 2006, using Refuge staff and/or local volunteers, randomly select one date (within the allowable tidally controlled time frame) in each eight day interval for upper route and one date in each 15 day interval for lower route to collect abbreviated, yet comparable data. Note that an increased interval between surveys increases the variance around arrival and peak date observations. 2) Discuss the survey results and methods with Migratory Bird Management biologists in Anchorage to see if they have recommendations to improve the survey. See if they would be willing to cooperate with funding or staffing a modified survey. 3) Discuss the disturbance survey methods with Migratory Bird Management to see if they can recommend improvements to the design. 4) Basic information about the Naknek staging waterfowl population needs to be compiled (i.e., major migration routes, length of staging period, wintering areas used by principal species, other nearby staging areas used when the Naknek River is unavailable, etc.). Establish cooperators on other refuges in the Pacific flyway where Naknek-staging waterfowl spend the winter and/or travel through in migration, in order to examine the relationships between survey data. Find out if any research has been done with regard to weather on those refuges in relation to migration phenology. ACKNOWLEDGMENTS Thanks to Bob Blush for providing training to the primary observer. Biological technician Kristin Sesser helped with several surveys and supported ACCESS data entry. Local observers Richard Russell and Rod Cyr provided additional helpful observations. Thanks to Refuge Manager Daryle Lons and Deputy Refuge Manager Mark Koepsel for providing support for the spring surveys. 18

24 LITERATURE CITED Banks, R.C. C. Cicero, J.L. Dunn, A.W. Kratter, P.C. Rasmussen, J.V. Remsen, J.D. Rising, D.F. Storz Forty-fifth supplement to the American Ornithologists Union Check-list of North American Birds. Auk 121(3): Bellrose, F.C Ducks, geese, and swans of North America. Third edition. Stackpole Books, Harrisburg, Pennsylvania. 540 pp. Ely, C.R. and Takekawa, J.Y Geographic variation in migratory behavior of greater white-fronted geese (Anser albifrons). Auk 113(4): Kessel, B. and D.D. Gibson Status and distribution of Alaska birds. Studies in Avian Biology, No. 1. R.J. Raitt and J.P. Thompson, eds. Cooper Ornithological Society, Los Angeles, CA. 100pp. Lapinski, S. and B. Williamson Spring staging waterfowl on the Naknek River, Alaska Peninsula, Alaska, March May Unpublished report, Alaska Peninsula/Becharof National Wildlife Refuge Complex, U.S. Fish and Wildlife Service, King Salmon, Alaska. MacGowan, B Spring staging of waterfowl along major drainages of Bristol Bay, with an emphasis on the Naknek River, Alaska March May Unpublished report, Alaska Peninsula/Becharof National Wildlife Refuge Complex, U.S. Fish and Wildlife Service, King Salmon, Alaska. Meixell, B.W. and S.E. Savage A survey of spring staging waterfowl along the Naknek River, Alaska Peninsula, Alaska, March-May 2002 and Unpublished report, Alaska Peninsula/Becharof National Wildlife Refuge Complex, U.S. Fish and Wildlife Service, King Salmon, Alaska. Russell, R. Personal communication. Local Naturalist and retired Alaska Department of Fish and Game Fishery Biologist, King Salmon, AK. Schuster, S.M Spring staging waterfowl on the Naknek River, Alaska Peninsula, Alaska, March May Unpublished report, Alaska Peninsula/Becharof National Wildlife Refuge Complex, U.S. Fish and Wildlife Service, King Salmon, Alaska. U.S. Fish and Wildlife Service (USFWS) Naknek Spring Waterfowl Census. Wildlife Inventory Plan. Alaska Peninsula/Becharof National Wildlife Refuge Complex, King Salmon, Alaska. Western Regional Climate Center < 19

25 Wilk, R Tundra swans in the Bristol Bay lowlands, Northern Alaska Peninsula. M.S. thesis, College of Natural Resources, University of Wisconsin. 20

26 2 TABLE 1. Summary of spring weather conditions in King Salmon, Alaska Peninsula, Alaska, with historical averages Year Ave MARCH Ave Temp (F) Tot. Precip (in) APRIL Ave Temp (F) Tot. Precip (in) MAY Ave Temp (F) Tot. Precip (in) year Ave. Ice-out date 6-Apr 25-Mar 30-Mar 29-Mar 7-Mar 8-Mar 9-Mar 18-Apr 31-Mar 31-Mar 17-Mar 9-Apr 5-Apr 1 7-Apr 2 29-Mar 3 28-Mar 1 First break-up was 12-Mar, first re-freeze 20-Mar, second break-up 25-Mar, second re-freeze 30-Mar First break-up was 23-Feb, first re-freeze 3-Apr. 3 First break-up was 19 March, first re-freeze 22 March. 21

27 TABLE 2. Date of first observation of principal species on the upper route, Naknek River, Alaska Peninsula, Alaska, Year Ave 2 SD (days) Survey beginning date (upper) 27-Mar 26-Feb 11-Mar 2-Mar 18-Mar 17-Mar 9-Mar 19-Mar 6-Mar 1-Mar 17-Mar 14-Mar 11-Mar 22-Mar 17-Mar Survey ending date (upper) 19-May 6-May 14-May 17-May 18-May 16-May 12-May 21-May 12-May 17-May 9-May 14-May 18-May 18-May 16-May Greater white-fronted goose 6-Apr 15-Apr 4-Apr 2-Apr 3-Apr 3-Apr 10-Apr 12-Apr 13-Apr 7-Apr 9-Apr 11-Apr 6-Apr 7-Apr 14-Apr 8-Apr 4.2 Canada / cackling goose 9-Apr 16-Apr 11-Apr 4-Apr 10-Apr 16-Apr 16-Apr 12-Apr 13-Apr 17-Apr 23-Apr 11-Apr 7-Apr 7-Apr 14-Apr 13-Apr 4.7 Tundra swan 27-Mar 16-Mar 29-Mar 23-Mar 25-Mar 17-Mar 24-Mar 8-Apr 6-Mar 13-Mar 29-Mar 27-Mar 17-Mar 22-Mar 25-Mar 23-Mar 7.8 Eurasian wigeon 21-Apr 1-Apr 31-Mar 9-Apr 10-Apr 16-Apr 10-Apr 17-Apr 3-Apr 10-Apr 15-Apr 15-Apr 14-Apr 7-Apr 24-Apr 12-Apr 7.0 American wigeon 15-Apr 2-Apr 13-Apr 9-Apr 28-Mar 16-Apr 22-Apr 17-Apr 13-Apr 5-Apr 9-Apr 11-Apr 7-Apr 22-Mar 21-Apr 10-Apr 8.5 Mallard 6-Apr 18-Mar 29-Mar 4-Apr 18-Mar 25-Mar 19-Mar 8-Apr 16-Mar 29-Mar 29-Mar 7-Apr 6-Apr 22-Mar 6-Apr 29-Mar 8.3 Northern shoveler 21-Apr 19-Apr 21-Apr 1-May 26-Apr 21-Apr 28-Apr 24-Apr 16-Apr 24-Apr 23-Apr 18-Apr 14-Apr 17-Apr 24-Apr 22-Apr 4.5 Northern pintail 2-Apr 11-Mar 29-Mar 2-Apr 25-Mar 3-Apr 22-Mar 17-Apr 20-Mar 13-Mar 9-Apr 7-Apr 17-Mar 22-Mar 6-Apr 28-Mar 10.6 Green-winged teal 16-Apr 2-Apr 31-Mar 4-Apr 12-Apr 16-Apr 22-Apr 19-Apr 16-Apr 19-Apr 29-Apr 18-Apr 14-Apr 17-Apr 24-Apr 15-Apr 8.0 Greater scaup 10-Apr 16-Mar 31-Mar 9-Apr 29-Mar 21-Mar 24-Mar 9-Apr 23-Mar 3-Apr 15-Apr 27-Mar 14-Apr 22-Mar 14-Apr 2-Apr 10.1 Common goldeneye 1 27-Mar 26-Feb 29-Mar 2-Mar 18-Mar 17-Mar 9-Mar 19-Mar 6-Mar 1-Mar 17-Mar 14-Mar 11-Mar 22-Mar 14-Mar NA Common merganser 1 27-Mar 26-Feb 11-Mar 2-Mar 18-Mar 17-Mar 9-Mar 19-Mar 6-Mar 1-Mar 17-Mar 14-Mar 11-Mar 22-Mar 14-Mar NA Red-breasted merganser 15-Apr 5-Apr 28-Mar 29-Mar 18-Mar 20-Mar 14-Apr 12-Apr 7-Mar 10-Apr 23-Apr 17-Mar 14-Apr 17-Apr 14-Apr 4-Apr resident species present on first day of survey 2 non-leap year calendar date used TABLE 3. Date of peak abundance of principal species on the upper route, Naknek River, Alaska Peninsula, Alaska, Year Ave 3 SD (days) Greater white-fronted goose 1-May 21-Apr 21-Apr 21-Apr 12-Apr 21-Apr 22-Apr 25-Apr 19-Apr 19-Apr 23-Apr 18-Apr 14-Apr 25-Apr 24-Apr 21-Apr 4.5 Canada / cackling goose 17-Apr 19-Apr 21-Apr 20-Apr 17-Apr 19-Apr 22-Apr 17-Apr 27-Apr 17-Apr 23-Apr 18-Apr 14-Apr 21-Apr 28-Apr 20-Apr 3.8 Tundra swan 1-May 21-Apr 21-Apr 21-Apr 1-May 21-Apr 23-Apr 25-Apr 27-Apr 17-May 9-May 29-Apr 16-Apr 17-Apr 28-Apr 27-Apr 8.3 Eurasian wigeon 1-May 26-Apr 27-Apr 1-May 3-May 30-Apr 6-May 13-May 19-Apr 10-May 9-May 25-Apr 28-Apr 25-Apr 28-Apr 1-May 6.4 American wigeon 1-May 6-May 27-Apr 1-May 8-May 28-Apr 10-May 11-May 3-May 10-May 9-May 5-May 30-Apr 4-May 28-Apr 4-May 4.8 Mallard 22-Apr 9-Apr 13-Apr 9-Apr 16-Apr 16-Apr 10-Apr 28-Apr 19-Apr 19-Apr 9-Apr 11-Apr 15-Apr 20-Apr 28-Apr 17-Apr 6.5 Northern shoveler 13-May 6-May 29-Apr 4-May 18-May 2-May 10-May 13-May 3-May 10-May 9-May 16-May 4-May 25-Apr 9-May 8-May 6.5 Northern pintail 1-May 26-Apr 22-Apr 24-Apr 27-Apr 21-Apr 23-Apr 11-May 27-Apr 8-May 9-May 25-Apr 18-Apr 23-Apr 24-Apr 28-Apr 6.9 Green-winged teal 13-May 3-May 27-Apr 1-May 29-Apr 21-Apr 6-May 11-May 27-Apr 10-May 29-Apr 29-Apr 28-Apr 25-Apr 9-May 2-May 6.5 Greater scaup 8-May 26-Apr 10-May 1-May 8-May 28-Apr 5-May 13-May 2-May 15-May 9-May 14-May 11-May 26-Apr 9-May 7-May 6.4 Common goldeneye 1 6-Apr 5-Apr 6-Apr 4-Apr 5-Apr 2-Apr 22-Mar 25-Apr 13-Apr 7-Apr 9-Apr 11-Apr 14-Apr 9-Apr 21-Apr 9-Apr 8.0 Common merganser 1 15-Apr 24-Mar 31-Mar 4-Apr 25-Mar 27-Mar 22-Mar 25-Apr 7-Mar 7-Apr 23-Mar 15-Apr 11-Mar 9-Apr 6-Apr 1-Apr 12.9 Red-breasted merganser 29-Apr 21-Apr 6-May 1-May 18-Mar 2-May 28-Apr 11-May 19-Apr 2-May 29-Apr 25-Apr 21-Apr 26-Apr 9-May 27-Apr resident species 2 peak abundance observed on two days (14,16-Apr) so average date of 15-Apr used in analysis 3 non-leap year calendar date used 22

28 TABLE 4. Peak abundance counts of principal species on the upper route, Naknek River, Alaska Peninsula, Alaska, Year Ave SD Greater white-fronted goose , ,188 1, ,425 1, Canada / cackling goose Tundra swan 2,649 1,343 1,719 1, ,177 1,291 1,161 1, ,284 2,220 2,679 1, Eurasian wigeon American wigeon Mallard Northern shoveler Northern pintail 11,480 2,610 6,321 1,730 1, ,974 2,979 3,723 4,750 2,425 3,130 5,302 3,618 5,315 3, Green-winged teal Greater scaup Common goldeneye , Common merganser 1 3,248 2,073 4,462 2,301 1,532 1,372 2,878 1,156 1,146 1,812 2, ,043 2,638 2,303 2, Red-breasted merganser Resident species 1 TABLE 5. Date of first observation, date of peak abundance, and peak abundance count of non-principal species on the upper route, Naknek River, Alaska Peninsula, Alaska, First Observation Peak Date Peak Count Brant 1 18-May 18-May 4 Trumpeter swan 2 14-Apr 14-Apr 4 Barrow's goldeneye 14-Mar 17-Mar 7 Canvasback 24-Apr 28-Apr 17 Bufflehead 28-Apr 28-Apr 6 Gadwall 9-May 9-May 3 1 Brant were seen on a disturbance survey 2 Although probably present, trumpeter swan were not separated from tundra swan on other surveys. 23

29 TABLE 6. Date of first observation of principal species on the lower route, Naknek River, Alaska Peninsula, Alaska, Year Ave 2 SD (days) Survey beginning date (lower) 24-Mar 1-Mar 24-Mar 3-Mar 20-Mar 17-Mar 11-Mar 14-Mar 15-Mar 2-Mar 28-Mar 18-Mar 12-Mar 17-Mar 15-Mar x x Survey ending date (lower) 29-May 5-May 13-May 12-May 14-May 15-May 14-May 27-May 12-May 16-May 10-May 22-May 20-May 17-May 15-May x x Mallard 14-Apr 30-Mar 7-Apr 7-Apr 26-Mar 18-Apr 7-Apr 26-Apr 12-Apr 2-Apr 19-Apr 9-Apr 12-Apr 6-Apr 7-Apr 10-Apr 7.9 Northern pintail 20-Apr 7-Apr 7-Apr 7-Apr 26-Mar 18-Apr 23-Mar 26-Apr 12-Apr 18-Apr 25-Apr 14-Apr 8-Apr 1 10-Apr 12-Apr 12-Apr 9.5 Greater scaup 14-Apr 7-Apr 25-Apr 26-Apr 10-Apr 29-Mar 28-Mar 18-Apr 12-Apr 7-Mar 28-Mar 28-Mar 10-Apr 14-Apr 7-Apr 7-Apr 12.9 Black scoter 29-Apr 7-Apr 4-Apr 5-Apr 18-Apr 18-Apr 19-Apr 21-Apr 12-Apr 9-Apr 25-Apr 22-Apr 23-Apr 17-Mar 26-Apr 15-Apr 11.4 Long-tailed duck 24-Mar 12-Mar 30-Mar 27-Mar 20-Mar 19-Mar 20-Mar 19-Mar 15-Mar 7-Mar 28-Mar 18-Mar 12-Mar 17-Mar 15-Mar 19-Mar 6.4 Common goldeneye 1-Apr 19-Mar 4-Apr 3-Mar 20-Mar 25-Mar 23-Mar 25-Mar 15-Mar 2-Mar 28-Mar 25-Mar 12-Mar 17-Mar 12-Apr NA Common merganser 24-Mar 1-Mar 24-Mar 3-Mar 20-Mar 17-Mar 11-Mar 14-Mar 15-Mar 2-Mar 28-Mar 18-Mar 12-Mar 17-Mar 15-Mar NA Red-breasted merganser 1-Apr 7-Apr 7-Apr 29-Mar 20-Mar 17-Mar 15-Apr 18-Apr 11-Apr 9-Apr 19-Apr 17-Apr 13-Apr 17-Apr 23-Mar 7-Apr Non-official survey 2 Non-leap year calendar date used TABLE 7. Date of peak abundance of principal species on the lower route, Naknek River, Alaska Peninsula, Alaska, Year Ave 2 SD (days) Mallard 23-Apr 25-Apr 25-Apr 21-May 15-Apr 22-Apr 15-Apr 26-Apr 12-Apr 18-Apr 25-Apr 17-Apr 13-Apr 14-Apr 26-Apr 22-Apr 9.4 Northern pintail 29-Apr 25-Apr 25-Apr 2-May 4-May 18-Apr 7-May 10-May 28-Apr 7-May 10-May 14-Apr 13-Apr 14-Apr 26-Apr 28-Apr 9.4 Greater scaup 14-May 25-Apr 29-Apr 2-May 25-Apr 8-May 14-May 10-May 12-May 16-May 3-May 15-May 19-Apr 3-May 10-May 6-May 8.3 Black scoter 29-May 8-Apr 14-Apr 12-May 9-May 15-May 14-May 27-May 4-May 16-May 10-May 7-May 10-May 17-May 15-May 10-May 13.6 Long-tailed duck 3-Apr 8-Apr 29-Apr 26-Apr 10-Apr 2-Apr 14-May 26-Apr 14-Apr 9-Apr 28-Mar 3-Apr 12-Apr 22-Apr 15-Mar 13-Apr 14.8 Common goldeneye 1-Apr 7-Apr 7-Apr 7-Apr 20-Mar 1-May 15-Apr 26-Apr 5-Apr 9-Apr 28-Mar 17-Apr 13-Apr 17-Mar 12-Apr 8-Apr 12.0 Common merganser 1-Apr 22-Mar 15-Apr 3-Apr 26-Mar 20-Mar 7-Apr 21-Mar 21-Mar 18-Mar 28-Mar 20-Mar 15-Mar 25-Mar 12-Apr 28-Mar 9.3 Red-breasted merganser 14-May 25-Apr 25-Apr 3-Apr 18-Apr 28-Mar 7-May 5-May 18-Apr 18-Apr 3-May 2-May 19-Apr 27-Apr 10-May 23-Apr Non-leap year calendar date used TABLE 8. Peak abundance counts of principal species on the lower Naknek River, Alaska Peninsula, Alaska, Year Ave SD Mallard Northern pintail Greater scaup Black scoter Long-tailed duck ,080 1, , , Common goldeneye Common merganser 3,270 1,639 1, , , , ,158 3,094 1, Red-breasted merganser

30 TABLE 9. Date of first observation, date of peak abundance, and peak abundance count of non-principal species on the lower route, Naknek River, Alaska Peninsula, Alaska, First Observation Peak Date Peak Count Canada Goose 26-Apr 26-Apr 2 Tundra swan 15-May 15-May 12 Eurasian wigeon 20-Apr 2-May 1 4 American wigeon 20-Apr 15-May 24 Northern shoveler 26-Apr 5-May 6 Green-winged teal 26-Apr 10-May 15 Canvasback 5-May 5-May 13 Bufflehead 5-May 5-May 14 1 Peak abundance observed on two days (20-Apr, and 15-May) so the average date of 2-May used in analysis. 25

31 Table 10. Summary of disturbance type frequencies during Naknek Waterfowl Ground and Disturbance Surveys, Alaska Peninsula, Alaska, General disturbance type Specific category No. on reg. survey No. on dist.-only survey Total Number % Total % Subtotals Avian Predators Bald Eagle - Adult Bald Eagle - Juvenile Common Raven Gulls Medium aircraft Aircraft - medium, twin prop Small aircraft Pedestrian Hunting Aircraft - medium, jet Aircraft - small, twin prop Aircraft - small, single prop Observer on Foot 0.0 Person on shore - conspicuous Shots fired (no people seen) Hunting (from shore/blind) Vehicles Vehicle - ATV Vehicle - Motor Boat <60' Vehicle - Truck/car Other Loud Noise Unknown Grand Total

32 Table 11. Summary of all disturbance categories and effect on waterfowl tribes present, Naknek Waterfowl Disturbance Survey, Naknek River, Alaska Peninsula, Alaska, Number of events is number that a member of that tribe is present for. Under Category Sum is sum of all birds responding to disturbance at that level. Category Scores: 1 = no effect, 2 = walk / swim away, 3 = flush and return within 200 m, 4 = flush and return > 200 m, 5 = Flush and leave area. GEESE SWAN Category Sum Category Sum Disturbance Source No. events No. events Aircraft - medium, jet Aircraft - medium, twin prop ,148 3 Aircraft - small, single prop ,568 2 Aircraft - small, twin prop Bald Eagle - Adult , Bald Eagle - Juvenile Common Raven Gulls Hunting (from shore/blind) Loud noise Person on shore - conspicuous Shots fired (no people seen) 38 8, , Unknown ,878 4 Vehicle - ATV Vehicle - Motor Boat <60' Vehicle - Truck/car Total Sum of No. Affected , ,

33 Table 11, con't. Summary of all disturbance categories and effect on waterfowl tribes present, Naknek Waterfowl Disturbance Survey, Naknek River, Alaska Peninsula, Alaska, Number of events is number that a member of that tribe is present for. Under Category Sum is sum of all birds responding to disturbance at that level. Category Scores: 1 = no effect, 2 = walk / swim away, 3 = flush and return within 200 m, 4 = flush and return > 200 m, 5 = Flush and leave area. DABBLER DIVER Category Sum Category Sum Disturbance Source No. events No. events 1 5 Aircraft - medium, jet Aircraft - medium, twin prop 5 4, Aircraft - small, single prop 14 6, Aircraft - small, twin prop Bald Eagle - Adult 15 4, Bald Eagle - Juvenile Common Raven Gulls Hunting (from shore/blind) Loud noise 1 1, Person on shore - conspicuous Shots fired (no people seen) 38 25, Unknown 11 4, Vehicle - ATV 2 1, Vehicle - Motor Boat <60' Vehicle - Truck/car 6 3, Total Sum of No. Affected , ,

34 Table 11, con't. Summary of all disturbance categories and effect on waterfowl tribes present, Naknek Waterfowl Disturbance Survey, Naknek River, Alaska Peninsula, Alaska, Number of events is number that a member of that tribe is present for. Under Category Sum is sum of all birds responding to disturbance at that level. Category Scores: 1 = no effect, 2 = walk / swim away, 3 = flush and return within 200 m, 4 = flush and return > 200 m, 5 = Flush and leave area. SEA DUCKS UNKNOWN Category Sum Category Sum Disturbance Source No. events No. events 1 5 Aircraft - medium, jet Aircraft - medium, twin prop ,550 Aircraft - small, single prop ,801 Aircraft - small, twin prop 1 96 Bald Eagle - Adult ,575 Bald Eagle - Juvenile 3 39 Common Raven ,500 Gulls ,500 Hunting (from shore/blind) Loud noise 1 16 Person on shore - conspicuous Shots fired (no people seen) 37 3, Unknown Vehicle - ATV Vehicle - Motor Boat <60' , Vehicle - Truck/car Total Sum of No. Affected , ,

35 Table 12. Disturbance rates, flushing rates (disturbance or disturbances causing flushing per hour), and flushing rates per disturbance by variable, Naknek Waterfowl Disturbance Survey, Alaska Peninsula, Alaska, Disturbance only survey data presented with means, standard deviations and results of statistical tests. P values less than 0.10 highlighted. Disturbance Rate Flushing Rate Flushes / Disturbance MONTH April Mean SD May Mean SD April vs. May t-test p-value LOCATION Paradise Point Mean SD Rapids Camp Mean SD PP Vs RC t-test p-value WEEKEND VS WEEKDAYWeekend Mean SD Weekday Mean SD WE vs WD t-test p-value PERIOD OF DAY Morning Mean no survey no survey no survey SD Midday Mean SD Afternoon Mean SD Evening Mean Mid-After-Eve SD ANOVA F 0.29 (Not 0.63 (Not statistic df (2,22) 0.05 (Not Sign) Sign) Sign) TIME OF DAY Hour vs Rate Correlation Coefficient

36 Table 13. Potential waterfowl disturbances per hour of observation by year, type of survey and major disturbance category, Naknek Waterfowl Disturbance Survey, Alaska Peninsula, Alaska, Ground Disturbance Ground Disturbance Ground Disturbance Ground Disturbance Bald Eagle Common Raven 0.16 Gulls 0.12 Aircraft - Large Aircraft - Medium Aircraft - Small Pedestrian (inc. Obs on foot) Shots Fired Hunting from (shore/blind) 0.02 Vehicles - ATV / Dirt Bike Motor Boat < 60' Motor Boat > 60' see above see above Vehicle - Truck / Car Misc Unknown Total

37 32

38 Figure 2. Map of the Naknek River Spring Waterfowl Lower Route Survey Points, Alaska Peninsula, Alaska. 33

39 Figure 3. Map of the Naknek River Spring Waterfowl Upper Route Survey Points, Alaska Peninsula, Alaska. 34