Movement Patterns ofHawaiian Petrels and Newell's Sheatwaters on the Island ofHawai'i1

Robert H. Day, 2 Brian A. Cooper, 3 and Richard]. Blaha3

Abstract: We studied movements and distribution and abundance of endan­ gered Hawaiian Petrels ('Ua'u [Pterodroma sandwichensis Ridgway]) and threat­ ened Newell's Shearwaters ('A'o [Puffinus auricularis newelli Henshaw]) on the island of Hawai'i in May-June 2001 and 2002. We recorded radar targets of either species at 14 of the 18 sites but recorded no visually at any site. Movement rates of petrels and shearwaters were very low (0-3.2 targets per hour) over all except one of the sites (Waipi'o Valley: 25.8 targets per hour). We saw radar targets moving from shortly after sunset throughout the rest of the sampling, suggesting that both petrels and shearwaters were present. Highest movement rates occurred 1-2 hr after sunset, when primarily Newell's Shear­ waters are flying. Timing of evening movements suggests that Hawaiian Petrels fly over the northern and southern parts of the island and may dominate on and . In contrast, timing suggests that Newell's Shear­ waters fly over essentially the entire island (except in the southwestern part, where no birds appear to occur), dominate numerically in the Moun­ tains, and occur in low numbers on Mauna Loa, in the Puna District, and on the northern slopes of Mauna Kea. Evening flight directions were predominantly inland at all sites except four. The limited radar data suggest that a substantial population change did not occur in the Puna District from 1995 to 2001-2002.

WITH THE EXCEPTION of the island of dearth of basic biological information is due Kaua'i, little information is available on primarily to these birds' nocturnal habits and the distribution, abundance, and population the inaccessibility of their remote, ­ trends of the endangered ous nesting colonies. Obtaining accurate in­ ('Ua'u [Pterodroma sandwichensis Ridgway]) formation on distribution, abundance, and and the threatened Newell's (Townsend's) population trends is an objective of the Re­ Shearwater ('A'o [Puffinus auricularis newelli covery Plan for both species (USFWS 1983) Henshaw]) in the . This and is especially important because of the re­ cent, precipitous declines of Newell's Shear­ waters on Kaua'i: in 2001, the number of shearwaters and petrels recorded on ornitho­ I The u.s. and Wildlife Service, Pacific Islands Ecoregion, , funded the 2001 research, and the logical radar around this island averaged only U.S. Coast Guard, Civil Engineering Unit, Honolulu, 38-40% of that recorded in June 1993 (Day funded the 2002 research. Manuscript accepted 14 Au­ et al., in press). gust 2002. The Hawaiian Petrel formerly was com­ 2 AER, Inc.-Environmental Research & Services, mon on the island of Hawai'i (Wilson and P.O. Box 80410, Fairbanks, Alaska 99708-0410 (phone: 907-455-6777; fax: 907-455-6781; E-mail: bday@abrinc. Evans 1890-1899). This seabird reportedly com). nested in large numbers on the slopes of 3 AER, Inc.-Environmental Research & Services, Mauna Loa, in the saddle between Mauna P.O. Box 249, Forest Grove, Oregon 97116-0249. Loa and Mauna Kea, and at moderate to high elevations on Hualalai (Wilson and Evans Pacific Science (2003), vol. 57, no. 2:147-159 1890-1899, Henshaw 1902, Richardson and © 2003 by University of Hawai'i Press Woodside 1954). For example, Munro saw All rights reserved one on the ground in 1891 at ~1370 m

147 148 PACIFIC SCIENCE· April 2003

(4500 ft) in Honaunau, Kona District (Banko (Sincock and Swedberg 1969, USFWS 1983, 1980a). The Hawaiian Petrel was a food Simons 1985). A secondary threat to these source of the ancient Hawaiians, and bones two species is their grounding (either through of this species are common in middens ex­ collision or exhaustion) and subsequent mor­ cavated in numerous locations on the island tality during the annual fallout of birds, pri­ of Hawai'i (Banko 1980a). By the beginning marily juvenile Newell's Shearwaters that are of the twentieth century, however, a decline on their way to the sea for the first time in this species had been noted by local resi­ (Hadley 1961, Telfer 1979, Sincock 1981, dents, and, by the early 1940s, Munro (1960) Reed et al. 1985, Telfer et al. 1987, Cooper feared for its survival in the Hawaiian Islands. and Day 1998, Podolsky et al. 1998, Ainley et However, this species still nests in low num­ al. 2001). Predation by introduced mammals, bers at higher elevations of Mauna Loa however, is thought to be the primary cause (Simons and Hodges 1998, Hu et al. 2001), of decline ofboth species on most islands and if not elsewhere on the island. is considered to be the major threat to the The Newell's Shearwater was thought to surviving populations at this time (USFWS be extinct in the Hawaiian Islands after 1894. 1983, Simons and Hodges 1998, Ainley et al. In 1954, however, a specimen was collected 2001). on O'ahu (King and Gould 1967), and a Ornithological radar, combined with visual breeding colony was found on Kaua'i in 1967 sampling, is useful for studying the move­ (Sincock and Swedberg 1969). Newell's ments and behaviors of these two species of Shearwater also breeds on the island of Ha­ nocturnal seabirds and is especially useful in wai'i, but probably in extremely low numbers: monitoring their populations. This research a few small nesting colonies recently have tool, which has been used in the Hawaiian been found in cinder cones in the Puna Dis­ Islands since 1992, has enabled much to be trict of eastern Hawai'i (Reynolds and Ri­ learned about the basic movements, behavior, chotte 1997). This species nests primarily in distribution, and/or population trends of burrows excavated under thick vegetation, these species around Kaua'i (Cooper and Day especially uluhe fern (Dicranopteris linearis 1995,1998, Day and Cooper 1995, Day et al., [Ainley et al. 1997]). in press), (Cooper and Day in press), Reasons for the declines of both species and Hawai'i (Reynolds et al. 1997; B.A.C. and are numerous. Introduced predators include R.H.D., unpubl. data). Most is known about wild pigs (Sus scrofa), Indian mongooses movements of these species on Kaua'i and (Herpestes auropunetatus), feral cats (Felis catus) Maui. and dogs (Canis familiaris), Barn Owls (Tyto Because of the recent declines of Newell's alba), Common Mynas (Acridotheres tristis) , Shearwaters on Kaua'i and the potential for and Norway (Rattus norvegicus), black (R. ­ continued development on all of the main tus), and Polynesian (R. exulans) , all of Hawaiian Islands, there is a clear need for which prey on both species, their eggs, and/ basic information on the distribution, abun­ or nestlings; in addition, introduced feral dance, and movement patterns of petrels and goats (Capra hircus) trample nesting colonies, shearwaters on all of the Hawaiian Islands. and endemic Short-eared Owls (Asio flam­ This study summarizes the results of radar meus sandwichensis) kill adults (Richardson and surveys for Hawaiian Petrels and Newell's Woodside 1954, Sincock and Swedberg 1969, Shearwaters on the island of Hawai'i during Byrd 1979, Byrd and Telfer 1980, Conant May-June 2001 and 2002. 1980, Byrd and Moriarty 1981, USFWS 1983, Simons 1984, 1985, Simons and Study Area Hodges 1998, Ainley et al. 2001, Hodges and Nagata 2001). Avian malaria and avian poxvi­ Hawai'i is the largest and easternmost of the ruses from introduced birds also may have main Hawaiian Islands. It is ~130 km (~81 negatively affected populations of both spe­ miles) in an east-west direction and ~ 145 km cies, particularly in low-elevation populations (~90 miles) in a north-south direction and Seabirds on Hawai'i . Day et at. 149

has a total area of ~1O,41O km 2 (~4021 ditions (daytime, crepuscular [twilight]' dark­ square miles). The island is dominated by ness), and moon phase/presence (i.e., whether four : the enormous Mauna Loa the moon was above or below the horizon). (4170 m [13,679 ft] high) in the southern part When a volunteer was available (11 of the 21 of the island, Mauna Kea (4206 m [13,796 ftl) nights), we also conducted visual sampling in the northeast, Hualalai (2522 m [8271 ftl) (lOx binoculars or 5x Noctron night-vision in the west, and the Kohala Mountains (1671 scope) for species identifications and flight m [5480 ftl) in the northwest. Mauna Loa, altitudes. Mauna Kea, and Hualalai are classic shield volcanoes, whereas the Kohala Mountains are Data Collection characterized by numerous steep valleys on their northern sides, much like northern We collected information on flying birds with Kaua'i. Habitats range from palm trees at sea a Furuno FCR-1510 MKIII (in 2001) or a level to rain forest (on windward slopes) or Furuno FCR-1411 (in 2002) surveillance ra­ xeric/mesic shrubs (on leeward slopes) at dar, which was an X-band radar transmitting moderate elevations and to snow and alpine at 9410 MHz with a peak power output of 12 vegetation near the summits of Mauna Loa kW (10 kW for the FCR-1411). The range of and Mauna Kea. this radar was set at 1.5 km (1.4 km for the FCR-1411), the pulse setting was 0.07 Ilsec (0.08 Ilsec for the FCR-1411), and the plot­ MATERIALS AND METHODS ting function was set to "continuous" (every We collected data on the movements of 15 sec for the FCR-1411). A similar surveil­ Hawaiian Petrels and Newell's Shearwaters lance radar is described in Cooper et al. at 15 sites between 31 May and 15 June 2001 (1991). The radar scanned a 3600 arc around and at three additional sites between 14 and the mobile radar laboratory and was used to 22 June 2002. All sites were situated on the obtain information on movement rates, flight perimeter of the island, as close as possible paths, and ground speeds of birds. This radar to the ocean. Sites were selected to pro­ has a digital color display with several scien­ vide both synoptic coverage of a region and tifically useful features, including color-coded information on individual locations (e.g., echoes (to differentiate the strength of return Waipi'o Valley) that were likely concentra­ signals), on-screen plotting of a sequence of tion points for inland flights of petrels and echoes (to depict flight paths), and True shearwaters. North correction for the display screen (to We sampled with ornithological radar for determine flight directions easily). (An echo 3 hr/night, between 1900 and 2200 hours in is a picture of a target on the radar display the evening; this part of the evening includes screen; a target is one or more birds displayed the peak of inland movement of these birds as a single echo on the radar display screen.) toward their nesting colonies (Day and The plotting function plotted the location of Cooper 1995). We attempted to collect data a target every sweep of the antenna; because during 25-min sessions, then used 5-min time intervals are fixed, ground speed is di­ breaks between periods of data collection to rectly proportional to the distance between collect weather data and to give observers a consecutive plots and can be measured with a short break. Actual lengths of sampling ses­ hand-held scale. We minimized ground clut­ sions were 14-25 min for radar data (some ter by elevating the forward edge of the time was lost when obscured the antenna, using a ground-clutter reduction radar screen), with 97 of 108 sampling ses­ screen mounted to the bottom of the antenna sions being the full 25 min long. We were face (described in Cooper et al. 1991), and unable to sample at all during two 25-min positioning the radar so that nearby vegeta­ sessions (one at Kona and one at Maulua tion acted as a radar fence (see Eastwood Stream). At the beginning of each sampling 1967). session, we recorded weather data, light con- For each echo seen on the radar display 150 PACIFIC SCIENCE· April 2003 screen, we recorded time, flight direction (to plete darkness (Day and Cooper 1995) to es­ the nearest 1°), and flight velocity (to the timate whether either species occurred there. nearest 8 lan/hr [5 milhr]). We collected data Hawaiian Petrels on Kaua'i head inland from only on targets flying ~48 lan/hr (~30 mil about sunset to a maximum of ~60 min after hr) and over land (following Day and Cooper sunset; however, most are moving up to about 1995, Day et al., in press). We also included the point of complete darkness (i.e., the end targets flying <48 lan/hr «30 milhr) that we of twilight), which occurs ~30 min after ­ identified visually as being of either of the set. In contrast, Newell's Shearwaters on two species of interest and excluded targets Kaua'i head inland from about 30 min after flying the appropriate speed but of another sunset (i.e., shortly after the point of com­ species; during this study, we added no petrel plete darkness) onward. Thus, the two species or shearwater targets flying <48 lan/hr and overlap in occurrence from the point of excluded no targets of other species that were complete darkness to ~30 min afterward. We flying the appropriate speed. Although we used these differences in timing on Kaua'i and attempted to collect information on species assumed no interisland differences in rela­ identification, flock size, and estimated flight tive timing to evaluate the species probably altitude of birds at 11 of the 18 sites, we had present at each site on Hawai'i. Probable no visual sightings of either species in 2001 or Hawaiian Petrels occurred between sunset 2002. and the point of complete darkness, whereas probable Newell's Shearwaters occurred from ~60 Data Analysis min after sunset onward. Possible Ha­ waiian Petrels or possible Newell's Shear­ All statistical analyses were conducted with waters occurred during this period of overlap SPSS or Microsoft Excel software. We tabu­ in movement, between ~30 and ~60 min lated numbers of targets recorded during after sunset (i.e., from the point of complete each sampling session, then converted these darkness to ~30 min later). Hence, we as­ counts to estimates of movement rates (tar­ sumed two competing scenarios for identity gets per hour), based on the number of during this period of overlap in movement: minutes sampled. Because rain showers (1) that all of the birds occurring during this sometimes obscured substantial portions of period of overlap were Hawaiian Petrels; and the screen, we subtracted that time from the (2) that all were Newell's Shearwaters. 25-min sampling period and used the result­ ing time in the calculation of movement rates. We used the estimated movement rates RESULTS for each sampling period to calculate the During our sampling, sunset occurred be­ mean ± 1 standard error (SE) movement rate tween 1854 and 1900 hours, or just before we for each night's sample and used the flight­ began sampling. Civil twilight ended (i.e., direction data to calculate the mean flight the point of complete darkness occurred) at direction for each night's sample. At the 1919-1924 hours, or ~25 min after sunset. Pa'auilo site in 2002, we had four consecutive Nightly variation in cloud cover and in to­ nights ofmovement data, so we calculated the pography and aspect of a site, however, re­ mean movement rate and mean flight direc­ sulted in a 10- to 15-min range among nights tion over the four nights at that site; other­ in the timing of the point of complete dark­ wise, all other sites had only one night of ness. In 2001, the sampling occurred from 2 data. To calculate flight directions, we con­ days after the first quarter moon to 2 days verted flight directions to radians and cal­ after the third quarter; hence, it occurred culated the mean direction following Zar during ~ 1 week on either side of the full (1984). moon. In 2002, the sampling occurred from At each site, we used data on the timing of midway between the new moon and the first movements of these two species on Kaua'i quarter to midway between the first quarter with respect to sunset and the point of com- and the full moon. Seabirds on Hawai'i . Day et at. 151

Pololu Valley Waipi'o Valley Pa'auilo Maulua Stream

Paradise Honokohau Park Hawaiian Beaches Kapoho saac Hale ehena

Hawa,.,-.

10 0 10 20 30 km EL;;;L"'''''~::''"'J The Main Hawaiian Islands

FIGURE 1. Geographic variation in movement rates (targets per hour) on surveillance radar around Hawai'i, May-June 2001 and 2002. Sizes of shaded circles are proportional to the rate of movement. White squares indicate location of towns.

Relative Abundance, Timing, and Species Identity (Waipi'o Valley). Further, targets were recorded at only 14 of the 18 sites, with none Very low numbers of petrels and shearwaters recorded at Isaac Hale, in southeastern Ha­ flew over all except one of the study sites wai'i, and at three sites in the southwestern (Figure 1). Mean nightly movement rates part of the island-Honokohau, Kona, and ranged from a to 25.8 targets per hour but Kealakekua. The highest movement rates exceeded 3.2 targets per hour at only one site were observed at Waipi'o Valley, which is 152 PACIFIC SCIENCE· April 2003

15 .. Point of complete darkness

12 n = 106 targets ::I: en-.. t- 9 W C) 0::: ~ 6

3

04--r--r--r--r---..,...-.,....--r--r--,...... ,r""'"""""I-,.--..,.---r---r..:::w~~ o 30 60 90 180150120

TIME AFTER SUNSET (MIN)

FIGURE 2. Evening pattern of movement of targets on surveillance radar at all study sites combined on Hawai'i, May­ June 2001 and 2002, by time after sunset. located on the wetter, windward coast of the 'Akaka Falls, Hilo, and Paradise Park), island. No birds were observed visually at any formed 100% of the targets (i.e., occurred of the sites. entirely during the period when only this Movement rates were variable within a species was moving) at one site (Maulua night, although radar targets moved from Stream), and may have formed 100% of the shortly after sunset throughout the rest of the targets (i.e., assuming that all of the targets night (Figure 2). In general, the highest rates recorded during the period of overlap were occurred ~1-2 hr after sunset, or ~30-90 of this species) at another three sites (Ke­ min after the point of complete darkness. A hena, Punalu'u, and Hilo). The occurrence of smaller pulse of inland-heading movement targets >60 min after sunset suggests that also occurred ~30 min after sunset, or just at Newell's Shearwaters occurred at 10 ofthe 14 the point of complete darkness. sites where we recorded radar targets (Ka­ Based on the timing of inland movements, poho, Holei Sea Arch, Ho'opl1loa, Kawaihae, proportions of the two species varied among Pololl1 Valley, Waipi'o Valley, Pa'auilo, sites. The occurrence of targets between 'Akaka Falls, Paradise Park, and Hawaiian sunset and the point of complete darkness Beaches), possibly occurred at another three (Day and Cooper 1995) suggests that Ha­ sites during the zone of overlap (Kehena, waiian Petrels occurred at 7 of the 14 sites Punalu'u, and Hilo), formed 100% of the where we recorded radar targets (Kehena, targets at four sites (Kapoho, Holei Sea Arch, Kawaihae, Pololl1 Valley, Maulua Stream, Pa'auilo, and Hawaiian Beaches), and may Seabirds on Hawai'i . Day et al. 153

have formed 100% of the targets (i.e., as­ data, Reynolds et al. 1997, Day et aI., in press) suming that all of the targets recorded during and the timing ofmovements observed in this the period of overlap were of this species) at study, we believe that most radar targets were another two sites (Ho'opuloa and Waipi'o Newell's Shearwaters, with smaller numbers Valley). The data strongly suggest that of Hawaiian Petrels. In the evening, Newell's Newell's Shearwaters were most abundant Shearwaters on Kaua'i tend to head into the in Waipi'o Valley and the Kohala Mountains colonies after it becomes completely dark in general. outside (Day and Cooper 1995, unpubl. data); Overall, the geographic pattern suggests similarly, nearly all of the birds observed in that the number of birds flying over the this study occurred after it became dark. southwestern part ofthe island (in the vicinity Likewise, all of the (few) targets seen visually of Hualalai) is so low as to approach zero. and identified to species in the Puna District The timing of evening movements suggests and in Waipi'o Valley in 1994 were Newell's that Dark-romped Petrels fly over the north­ Shearwaters (Reynolds et al. 1997). Hence, ern and southern parts of the island and may most of the radar targets we observed at our dominate on Mauna Loa and Mauna Kea. In study sites in 2001-2002 probably were contrast, the timing suggests that Newell's Newell's Shearwaters, although it is likely Shearwaters fly over essentially the entire is­ that at least a few Hawaiian Petrels also flew land (except in the southwestern part), domi­ over our sites, based on the predarkness tim­ nate numerically in the Kohala Mountains, ing of movements of birds at several sites and and occur in low numbers on Mauna Loa, in the fact that small numbers breed on Mauna the Puna District, and on the northern slopes Loa (Simons and Hodges 1998, Hu et al. ofMauna Kea. 2001).

Flight Direction and Location Distribution and Abundance Evening flight directions were predominantly The low numbers of seabirds that we re­ inland at all sites except Holei Sea Arch, corded flying inland at all sites except Waipi'o Ho'opuloa, Pa'auilo, and 'Akaka Falls (Figure Valley match well with the limited informa­ 3). In this figure, we have excluded sites tion that is available on the distribution of where we recorded less than three flight di­ Hawaiian Petrels and Newell's Shearwaters rections. At Ho'opuloa and Ho1ei Sea Arch, on the island of Hawai'i (Ainley et al. 1997, the few targets that we did record were Reynolds et al. 1997, Simons and Hodges heading seaward, but the mean flight direc­ 1998; R.H.D. and BAC., unpubl. data). tion paralleling the coast at Pa'auilo and With the exception of Waipi'o Valley, some 'Akaka Falls was an artifact of peculiarities of cinder cones in the Puna District of eastern circular statistics and small sampling sizes Hawai'i, and some small colonies high on (some targets heading seaward and some Mauna Loa and probably Mauna Kea, petrels heading inland, resulting in a directional vec­ and shearwaters are thought to be scarce on tor that was nearly at right angles to those this island. directions). The distribution of the Hawaiian Petrel on the island of Hawai'i is imperfectly known but seems to match the pattern we saw in DISCUSSION 2001-2002. We can find no definite records Species Identity of this species within the Puna District for at least the last 30 yr, although our data for Although we were unable to observe any of Paradise Park and Kehena suggest that a few the birds visually to determine their identity, fly over this area on their way inland. In the based on our previous studies on Kaua'i, Ka'u District, the species has been reported Maui, and Hawai'i (Cooper and Day 1995, repeatedly in Hawaiian Volcanoes National 1998, in press, Day and Cooper 1995, unpubl. Park (van Riper and Barbee 1978, Banko 154 PACIFIC SCIENCE· April 2003

Pololu Valley

Paradise Park

Ho'opuloa

Hawal';~ 10 0 10 20 30 km ~~Wiii;;-rtr::'??!?1

The Main Hawaiian Islands

FIGURE 3. Mean flight direction of targets observed on radar around Hawai'i, May-June 2001 and 2002. Only sites with three or more targets are shown. White squares indicate location of towns.

1980a, Pyle 1984, 1986c) and on Mauna Loa corded only a few possible petrels flying (Banko 1980a, Conant 1980, Pyle 1992b, Hu inland at Ho'opuloa but saw none at Keala­ et a1. 2001; B.A.C. and R.H.D., unpub1. data); kekua, suggesting that most reach Mauna Loa our data indicate that this species probably from a direction other than the southwest or flew over Holei Sea Arch and Punalu'u. We west. In the North Kona District, the only can find no recent records of Hawaiian Pe­ recent record of this species that we can find trels from the South Kona District; we re- is of calling birds on Hualalai in the 1970s Seabirds on Hawai'i . Day et at. 155

(Conant 1980); because a subsequent VISIt (Kepler et al. 1979); however, Banko (1980b) there failed to turn up any birds, the few that recorded none afterward, and the recent were recorded may have been passing farther records on ornithological radar from the inland to nesting areas on Mauna Loa. Simi­ national park (Reynolds et al. 1997) are of lar to this apparent rarity, we recorded no questionable accuracy (see later in this dis­ petrels (or shearwaters) at Kona or Honoko­ cussion). Elsewhere in the Ka'l1 District, hau. There also are no recent records of Ha­ these shearwaters are believed to nest at waiian Petrels from either the South Kohala moderate elevations on the southeastern or North Kohala (plus Waipi'o Valley) Dis­ slopes of Mauna Loa (R. E. David, Kailua­ tricts (i.e., the Kohala Mountains); however, Kona, pers. comm.), a belief strengthened by our data for Kawaihae and Pololl1 Valley and our data from Holei Sea Arch and possibly probably Waipi'o Valley suggest that petrels Punalu'u; however, BAC. (unpubl. data) did nest somewhere in those mountains in low not record them visually in the Pohakuloa numbers. The species formerly nested, and Training Area or farther up on Mauna Loa in probably still does, somewhere on Mauna 1995 (no radar sampling was possible because Kea, given the number ofboth old and recent of the presence of large numbers of noctur­ records from there and the and nal insects), suggesting that, if they nest on North Hilo Districts (Richardson and Mauna Loa, they do so only in low numbers Woodside 1954, Banko 1980a, Pyle 1986a,b, and/or at moderate elevations on that moun­ 1992a); likewise, our data indicate that petrels tain's steep southeastern slopes. In the South fly over both Maulua Stream and 'Akaka Falls Kona District, the only records we can locate on their way inland. We found historical are a specimen and several records of (Baldwin and Hubbard 1949), but no recent, Newell's Shearwaters heard calling at night records ofHawaiian Petrels in the South Hilo as they flew over the Papa area between 1970 District; however, our data indicate that the and 1972 (Banko 1980b); our data from species passes over the Hilo area while head­ Ho'opl1loa indicate that Newell's Shearwaters ing inland. still use this area, although their absence from The Newell's Shearwater has been re­ Kealakekua suggests that they are rare there. corded in numerous locations on the island of We can find no recent or historical records of Hawai'i, although information on exact nest­ this species from the North Kona District, ing locations is limited. In the Puna District, and our data suggest that the species is absent the species recently has been found to nest in from or occurs in very low densities in that at least one (Reynolds and Ri­ district. We also can find no recent or his­ chotte 1997), and numerous visual and radar torical records of this species from the South records (Pyle 1986a, 1992b, Reynolds et al. Kohala District, but, in the North Kohala 1997; R.H.D. and BAC., unpubl. data) sug­ District (plus Waipi'o Valley), the species gest regular, albeit low-level, use of the area. must nest in at least one location, and proba­ The one exception was a stream of Newell's bly more, in the Kohala Mountains; in par­ Shearwaters seen flying inland in the "Kala­ ticular, the species must nest in substantial pana district" over an "extended period" in numbers somewhere up Waipi'o Valley, as the summer of 1992 (Pyle 1992b). In this indicated by the large number of visual (Hall district, we recorded this species at Paradise 1978, Kepler et al. 1979, Reynolds et al. 1997) Park, Hawaiian Beaches, and Kapoho and and radar (Reynolds et al. 1997; this study) possibly at Kehena (some targets occurred records from that location. The species also during the zone of overlap of the two spe­ must nest somewhere on the northern slopes cies), but not at Isaac Hale, suggesting pres­ of Mauna Kea, given the substantial number ence but low overall abundance. In the Ka'l1 of records from the Hamakua, North Hilo, District, information on the species' distribu­ and South Hilo Districts (Ralph and Pyle tion and abundance is unclear. The species 1977, Kepler et al. 1979, Banko 1980b, Con­ was recorded breeding in Makaopuhi Crater ant 1980, Pyle 1990, 1998); similarly, our data (southeast of Kilauea ) in 1972 suggest that this species occurred at Pa'auilo, 156 PACIFIC SCIENCE· April 2003

'Akaka Falls, and Hilo. In this part of the sual identifications of birds while sampling at island, the most promising additional sit:.es for that site in 2001; however, Reynolds et al. this species include the vicinity of 'Awehi (1997) did visually identify four Newell's Stream, Ka'uku Hill, Hakalau Stream, and Shearwaters and one unidentified shearwater/ the . petrel, but no Black Noddies or Sooty Terns, at Waipi'o Valley in 1994, indicating that at Population Trends least some of their targets were correctly identified. Although our overall pattern of abundance In contrast to the pattern seen for the 1994 was similar to that from other studies, the Reynolds et al. (1997) study, July 1995 data movement rates that we observed at Waipi'o both from and near the Kehena site averaged Valley and in the Puna District of eastern ~0.4 targets per hour across four nights and Hawai'i in 2001-2002 were lower than those mornings of sampling (BAC. and R. E. observed by Reynolds et al. (1997) in 1994. In David, unpubl. data), or slightly lower than contrast, our results in 2001-2002 were simi­ what we saw in this study. These data are lar to those in a study at Kehena in 1995 more comparable (i.e., collected and analyzed (BAC. and R. E. David, unpubl. data) and a similarly) with those from this study and sug­ study in the Puna District in 2001 (unpubl. gest that there has not been a substantial data). decline in the number of petrels and shear­ We conclude that the Reynolds et al. waters visiting the Puna District between (1997) estimates are not comparable with 1995 and 2001-2002. those in this or any other study and that the During fall 2001, we (unpubl. data) also identity of some of their radar targets should collected and analyzed data in the Puna Dis­ be considered questionable. The data upon trict similarly to what we did in this study. which the 1994 estimates were based are ex­ We recorded means of 1.5 targets per hour at tremely limited at best, they were collected Kehena, 0.8 targets per hour at Isaac Hale, only during the peak of movement (inflating 0.4 targets per hour at Kapoho, 0.4 targets estimates of mean movement rates over an per hour at Hawaiian Beaches, and 1.2 targets evening), some were collected during a period per hour at Paradise Park. Hence, these re­ (morning) in which the estimates would be cent fall data also are similar to the pattern inflated because that period of peak move­ reported here for the Puna District in sum­ ment is shorter and more intense than that in mer 2001-2002. the evening, and some of the data almost certainly were of insects (especially at places such as Pali Uli), rather than birds, or were Comparison with Other Islands birds of other species. In the study reported here, some insect contamination also is pos­ This study indicates that the number of sible for the Waipi'o Valley site, where we Hawaiian Petrels and Newell's Shearwaters observed hundreds of targets in 2001 flying visiting the island of Hawai'i is fairly small, just below the 30 miles/hr cutoff speed. It is probably on the order of a few thousand possible that these were petrels or shear­ birds. At most locations we sampled, move­ waters flying slowly against a headwind that ment rates were low, ranging between 0 and was not apparent on the ground (on the ~3 targets per hour, suggesting that 0-10 ground, we recorded an easterly tailwind, birds flew over each site in the first 3 hr ofthe which would have increased the velocity of evening. Obviously, the number of birds vis­ organisms flying westward), but our best iting the Kohala Mountains is considerably guess is that these slow targets primarily were larger, with probably several hundred flying large moths traveling with a tailwind and/or up Waipi'o Valley and the two similar valleys smaller seabirds such as Black Noddies (Anous immediately to the west of it and others minutus) or Sooty Terns (Sterna fuscata). Un­ coming in along the rest of the coastline in fortunately, we were unable to obtain any vi- that area. Seabirds on Hawai'i . Day et al. 157

Mean evening movement rates at allioca­ Banko, W. E. 1980a. Part I. Population tions on Hawai'i except Waipi'o Valley were histories-species accounts. Sea birds: much lower than those at all sites examined Hawaiian Dark-romped Petrel ('Ua'u). on Kaua'i (Day and Cooper 1995, Day et al., Cooperative National Park Resources in press) and were comparable to those at the Studies Unit, University of Hawai'i at lowest-movement sites on southwestern Maui Manoa, Honolulu. CPSU/UH Avian His­ (Cooper and Day in press). Clearly, however, tory Report 5B: History of Endemic Ha­ the mean movement rate at Waipi'o Valley waiian Birds. was comparable to some on the eastern shore ---. 1980b. Part 1. Population histories­ of Kaua'i and with moderate-movement sites species accounts. Sea birds: Newell Shear­ on Maui. The pattern of higher movement water ('A'o). Cooperative National Park rates on the wetter, windward side of Hawai'i Resources Studies Unit, University of Ha­ in this study follows those from these other wai'i at Manoa, Honolulu. CPSU/UH two islands, where rates also are higher on the Avian History Report 5B: History of En­ wetter, windward sides and lower on the demic Hawaiian Birds. drier, leeward sides of the islands. Byrd, G. V. 1979. Common Myna predation on Wedge-tailed Shearwater eggs. 'Ele­ paio 39:69-70. ACKNOWLEDGMENTS Byrd, G. v., and D. Moriarty. 1981. Treated Colleen Henson (U.S. Fish and Wildlife chicken eggs reduce predation on shear­ Service) and Jay Silberman (U.S. Coast water eggs. 'Elepaio 41:13-15. Guard) administered the study. Valerie Byrd, G. v., and T. C. Telfer. 1980. Barn Graber, Darcy Hu, and Roberta Swift helped Owls prey on birds in . 'Elepaio with visual observations. At ABR, we thank 41:35-36. Tai Graham for providing the satellite map Conant, S. 1980. Recent records of the image and Allison Zusi-Cobb for graphics 'Ua'u (Dark-romped Petrel) and the 'A'o advice. This article was improved by the ef­ (Newell's Shearwater) in Hawaii. 'Elepaio forts of two anonymous reviewers. 41:11-13. Cooper, B. A., and R. H. Day. 1995. Literature Cited endangered seabird study. Vol. 1. Inter­ actions of Dark-romped Petrels and Ainley, D. G., T. C. Telfer, and M. H. Rey­ Newell's Shearwaters with utility struc­ nolds. 1997. Townsend's Shearwater (pzif­ tures on Kauai, Hawaii. Electric Power finus auricularis). In A. Poole and F. Gill, Research Institute, Palo Alto, California, eds. The birds ofNorth America, No. 297. Report No. TR-I05847-V1. Academy of Natural Sciences, Philadel­ ---. 1998. Summer behavior and mortal­ phia, Pennsylvania, and American Orni­ ity of Dark-romped Petrels and Newell's thologists' Union, Washington, D.C. Shearwaters at power lines on Kauai. Co­ Ainley, D. G., R. Podolsky, L. de Forest, G. lon. Waterbirds 21: 11-19. Spencer, and N. Nur. 2001. Status and ---. In press. Movement of Hawaiian population trends of the Newell's Shear­ Petrels to Inland Breeding Sites on Maui water on Kauai: A model for threatened Island, Hawaii. Waterbirds. petrels on urbanized tropical oceanic is­ Cooper, B. A., R. H. Day, R. J. Ritchie, and lands. Pages 108-123 in J. M. Scott, S. C. L. Cranor. 1991. An improved marine Conant, and C. van Riper III, eds. Ecol­ radar system for studies of bird migration. ogy, conservation, and management of J. Field Ornithol. 62:367-377. endemic Hawaiian birds: A vanishing avi­ Day, R. H., and B. A. Cooper. 1995. Patterns fauna. Stud. Avian Biol. 22. of movement of Dark-romped Petrels and Baldwin, P. H., and D. H. Hubbard. 1949. Newell's Shearwaters on Kauai. Condor The Dark-romped Petrel reappears on 97:1011-1027. Hawaii. Condor 51:231-232. Day, R. H., B. A. Cooper, and T. C. Telfer. 158 PACIFIC SCIENCE· April 2003

In press. Decline of Newell's Shearwaters 1986c. Hawaiian Islands Region on Kauai, Hawaii. Auk. (1 August-30 November 1985). Am. Birds Eastwood, E. 1967. Radar ornithology. Me­ 40:161-163. thuen, London, U. K. ---. 1990. Hawaiian Islands Region (1 Hadley, T. H. 1961. Shearwater calamity on August-30 November 1989). Am. Birds Kauai. 'Elepaio 21 :60. 44:166-168. Hall, J. B. 1978. Newell Shearwater in the ---. 1992a. Hawaiian Islands Region Kohala Mountains, island of Hawaii. 'Ele­ (winter 1991-1992). Am. Birds 46:317­ paio 38:131-132. 318. Henshaw, H. W. 1902. Complete list of birds ---. 1992b. Hawaiian Islands Region of the Hawaiian Possessions with notes on (summer 1992). Am. Birds 46:1181-1182. their habits. Thrum, Honolulu, Hawai'i. ---. 1998. Hawaiian Islands Region (fall Hodges, C. S. N., and R J. Nagata. 2001. migration 1997). Field Notes 52:130-131. Effects of predator control on the survival Ralph, C. J., and R L. Pyle. 1977. Hawaiian and breeding success of the endangered Islands Region (1 June-31 July 1977). Am. Hawaiian Dark-romped Petrel. Pages Birds 31:1192-1193. 308-318 inJ. M. Scott, S. Conant, and C. Reed, J. R, J. L. Sincock, and J. P. Hailman. van Riper III, eds. Ecology, conservation, 1985. Light attraction in endangered pro­ and management of endemic Hawaiian cellariiform birds: Reduction by shielding birds: A vanishing avifauna. Stud. Avian upward radiation. Auk 102:377-383. BioI. 22. Reynolds, M. H., and G. L. Richotte. 1997. Hu, D., C. Glidden, J. S. Lippert, L. Schnell, Evidence of Newell's Shearwater breeding J. S. MacIvor, and J. Meisler. 2001. Habi­ in Puna District, Hawaii. J. Field Ornithol. tat use and limiting factors in a population 68:26-32. of Hawaiian Dark-romped Petrels on Reynolds, M. H., B. A. Cooper, and R H. Mauna Loa, Hawai'i. Pages 234-242 in J. Day. 1997. Radar study of seabirds and M. Scott, S. Conant, and C. van Riper III, bats on windward Hawai'i. Pac. Sci. 51:97­ eds. Ecology, conservation, and manage­ 106. ment of endemic Hawaiian birds: A van­ Richardson, F., and D. H. Woodside. 1954. ishing avifauna. Stud. Avian BioI. 22. Rediscovery of the nesting of the Dark­ Kepler, C. B., J. Jeffrey, andJ. M. Scott. 1979. romped Petrel in the Hawaiian Islands. Possible breeding colonies of Manx Condor 56:323-327. Shearwater on the island of Hawaii. 'Ele­ Simons, T. R 1984. A population model of paio 39:115-116. the endangered Hawaiian Dark-romped King, W. B., and P. J. Gould. 1967. The sta­ Petrel. J. Wildl. Manage. 48:1065-1076. tus of Newell's race of the Manx Shear­ ---. 1985. Biology and behavior of the water. Living Bird 6:163-186. endangered Hawaiian Dark-romped Pet­ Munro, G. C. 1960. Birds ofHawaii. Rev. ed. rel. Condor 87:229-245. C. E. Tuttle Co., Rutland, Vermont. Simons, T. R, and C. N. Hodges. 1998. Podolsky, R, D. G. Ainley, G. Spencer, L. de Dark-romped Petrel (Pterodroma phaeopy­ Forest, and N. Nur. 1998. Mortality of gia). In A. Poole and F. Gill, eds. The birds Newell's Shearwater caused by collisions of North America, No. 345. Academy of with urban structures on Kauai. Colon. Natural Sciences, Philadelphia, Penn­ Waterbirds 21:20-34. sylvania, and American Ornithologists' Pyle, R L. 1984. Hawaiian Islands Region Union, Washington, D.C. (1 August-30 November 1984). Am. Birds Sincock, J. L. 1981. Saving the Newell's 39:105-107. Shearwater. Pages 76-78 in Proceedings of ---. 1986a. Recent observations, March­ the Hawaii Forestry and Wildlife Confer­ June 1986. 'Elepaio 46:156-157. ence, 2-4 October 1980. Department of ---. 1986b. Recent observations, June­ Land and Natural Resources, State of Ha­ July 1986. 'Elepaio 46: 175-176. wai'i, Honolulu. Seabirds on Hawai'i . Day et at. 159

Sincock, ]. L., and G. E. Swedberg. 1969. 1983. Hawaiian Dark-rumped Petrel and Rediscovery of nesting grounds of Newell's Manx Shearwater Recovery Plan. Newell's Manx Shearwater (Puffinus puffi­ US. Fish and Wildlife Service, Portland, nus newelli), with initial observations. Oregon. Condor 71:69-71. van Riper, c., III, and R. D. Barbee. 1978. Telfer, T. C. 1979. Successful Newell's Shear­ Dark-rumped Petrel at Hawaii Volcanoes water salvage on Kauai. 'Elepaio 39:71. National Park. Bird-Banding 49:372. Telfer, T. c.,]. L. Sincock, G. V. Byrd, and Wilson, S. B., and A. H. Evans. 1890-1899. ]. R. Reed. 1987. Attraction of Hawaiian Aves Hawaiiensis: The birds of the Sand­ seabirds to lights: Conservation efforts and wich Islands. R. H. Porter, London, U K. effects of moon phase. Wildl. Soc. Bull. Zar,]. H. 1984. Biostatistical analysis. 2nd ed. 15:406-413. Prentice-Hall, Englewood Cliffs, New USFWS (US. Fish and Wildlife Service). Jersey.