Richard E. Filzner ano William H. Rickard Ecosystems Department Pacific North* est Laboratory P.O.Box 999 Richland,\ifashingron 99152

Canada Goose NestingPerformance along the Hanlord Reach of the ColumbiaRiver, 1971 -1981

Abslracl Clutch sizc obscrvcdin 1215 hatchcdGreat Baiin CanadaGoose (Brunta canaden:ismoliti) n.sts from 1971'1981was 5.6.The averagepercent of succcssfulnests was 82 percent.A meano{ 1J5 nesrs per year since1971 is lower than the 215 nestsper yearobserved during 1951'1970. Lossof nestins due to coyotepredation on Lxke Island-#6 has accountedfor the overall drop in number of nests. lntroduclion The Hanford nesting population of the Canada Goose has been studied since 1950. Han- soo and lberhardt (1971) have discussed the 1913-1970 period io great detail. This report examines data collected from 1971 ro 1981 and continues a recotd of an importanc nesting population of the Great Basin Canada Goose. One of rhe inirial purposes of these studies rvas to documeot the reproductive perfotmance of the goose population, and this aspect of the investigation has continued to determine whether nesting lxrformance *'ould demonstrate a delayed response to nuclear reactor operatioos. Radionuclide con- tent of C,anadagoose eggs measured after the closure of the production reactors indi- cated that the radionuclide content of goose ellgs aken ftom deserted nests along the w'as low and primarily of *.orld*ide fallout origin (Rickard and Sweany 1977) . ContintLous documentation of nesring performance alrc provides a way to evaluate the effect of future industrial uses of q'ater and any habitat changes ioduced by hyclroelectric dams ancl turbine additions up and dou'ostrsam from the Han- "park' ford reach. The progrsed establishment of a comtrercial nuclear power reactor at the Haolord Site to produce electriciry for export to the regional po*'er network could also produce a numt'er of environmental changes that could affect the nesring goose population as well as other wildiife populations. A hydroelectric dam across the Hanforcl reach would inundate the islands upo.n rvhich the nesting goose population depends. The recent opeoing of the entite leogth of the Haoford Reach to public recreational use is another f€arure that could have deleterious effect on the nesting geese. Sequentially col- lected data can serve as a way to evaluate the effects of past environmental changes and perhaps to recommeod fuure mitigation practices to help maintaio a dimioishing wildlife resource in south-central lfashinston.

Sludy Area The Hanford Reach (Figure 1) contains the only free-flowing grrtion of the Columbia

Northtest Science,Vol. 57,No.4, 19lJl 261 I N I PRIEST ' RAPIDS RIVERFLOW DAM

HANFORDSIiE

a REACTORAREAS ISLANDNUMBER RTSERVATIONBOUNDARY RIC HLAND

YAKIMA RIVER

Figure 1. The Hanford reach of the Columbia River, showing locations of 20 isiands used fot nesting by Canada Geese.

River in the upstteam from Bonneville Dam. The riparian and aquatic habitatsrepresent relatively unmanagedecological resources. Twenty islands provide almost all of the goosenesting habitats along the Hanford Reach (Hanson and Eberhardt 197i). Someof theseislands have changedvegeratively since 1970, particulady Islands 18, I), ar'd 20, as a result of pool elevationsof Lake Wallulla, the irnpoundmentcreated by McNary Dam in 1955. Here the establishmentof tree aod shr.ub willows (.Sallr spp, and rank herbaceous species rcxtted io soil and mud substrates; e.g., reed canary grass (Phalarit arund,i,nacea)) have replaced the sparse, sholr-sraturedplant communitiesrooted in cobblesrones aod gravelswhich were adapted to rhe historical seasonalflooding regime and the rapid flows of the free-flowing Colum- bia River (Fickeiseo et al. 1)80). Lcrke Island (#6, Fig. 1) has also been grazed by a small resident herd of feral cattle,and Island #il was burned by human cafelessoess.

Methods Nesring surveysduring i971-1!81 were conductedbiweekly and usually begaa during the first rveekof April, as describedby Hanson and Eberhardt (1971). Prior to 1971, nesting surveyswere cooducredweekly, thus not all prarametersmeasurd in the earlier

268 Firzner and Rickard surveys can be compated to the post-1970 data base. Nfe feel, however, thar rhe nest per- forrnance parameters n'e have selected are generaily comparable ro rhose of earlier iovesri- gations and serve ro exreod the period of ol-rservation ao additional 10 years. Certain pafametcrs. such as nesting phenology, nestiog success,and fates of eggs, are sensirive to the frequency of observation periods and probably differ slightly frorn pre-1970 data.

Resultsand Discussion The anoual number of goose nests on each island cao be an effective measure of the changing starus of the population. Horvever, other parameters such as nesr successj number of eges laid in successful oests, and avetage clutch sizes aiso provide r:seful ..nestiog information (Table 1) on the reproductive porential. The parameter success', as s'e use it, means rhar a flesr \aas successful if it contained the shells and membrarres of hatched egE1s,filoplumes from natai dorvn, pipped eggs, or goslings (Hanson and Ibedrardt 1971). Average clutch size is rhe average clutches from all successful nests. I(e compare productiviry dara io Table 1 with Haoson aod EberhardCs (1971) fig,ures lor the saore patameters. In 121j hatched nests from 1971-19S1 we obsefted a mean clutch size of 5.6. The avera.qepetcent of successful nests reported by Hanson and Eber- hardt was 71 percent, while our value u,as 82 percent. The generally higher values we report for successprobably refiect the biweekly frequency of oest checks. The potential for the goose grpulation to teach rhe 1958 level is apparently still presenr based on pro- ductivity parameters.

'l'-{ELTI L hoduclilii,r of Crudr coose nes|s on ihe Ilaniord }t!{cl oi the Cohunbu llivcr. 1lri1 io 1!t0.

1971 1!i2 19?r 1!7,1 1!?, 1!?6 19?? 1!73 19?9 trS0 1931 Ars. 112 I60 l:i t.t li 108 111 125 1.11 136 1t6 160 135

jl successlul rcsts 1l1l 13 I 1ii 111 !s 100 if 112 11]] 111 114 110

succ,.ssiul n.sts 35 "!il !t iti !1 110 62 i! sl 7l sl s,

succ-.ssilrl nesls 566 i41 6.13 5S2 t23 Srij |ia 61.1 6.1iJ 607 E83 620 -avcraee dutch sjze 5.1 5.6 i.a n.2 5..1 5.,1 t.S t.8 5.7 6.6 6.1 li.6

The number of gooseoesrs esrablished on the Haoford reach islandshas flucruated trom year to year, but a generaldecline in overail oumbers is evident. More than 300 goosenesrs \r-ere presenr in 1958, but io i975 only 108 oestsu,ere couoted (Fig. 2). This observeddeciine in rhe numbel of nesring atempts! we believe,is due to a combi- nation of facrors,with the commoo (Cdni! latrans) assrgnedan impotant role in the decline. The displacementof a residenthuman lx.rpulationfrom the reach of the Columbia in 1!43 *'as inirially beneficial to the goosepopulation, by reducing human visirationsto the islands-This kind of site managementalso benefitted the coyotepopu- lation by provicling a releasefrom conttol rneasures. The most dtamatic effect of rhe coyoreon the goose popularion is iilustrated for Locke Island (Island #6), which has 106 acresaod is the largestof the islands (Fig. 3). In 1957,Locke Island supportedat leasr 129 goosenesrs (Hanson and Eberhardt 1971). The first evidenceof coyote invasion of the island rvas recordedio 1959. rvheo 42 of 96 nests were destroyed (Hanson aod Eberhardt 1971). The secondcoyote incident

CanadaGoose Nesting oo the Columbia,1971-1981 269 54 56 58 60 6? 64 66 68 70 t2 14 16 YEAR

Iigure 2- Numbcr of Canadacoose nestsfound on rhe Hanford Reach,1954-1981.

2 succtssrw E rornr-

54 56 58 60 62 64 66 68 70 72 74 76 78 80 YIARS

Figure 3. Number of C-anadaGoose nests established on Island 6 (Iocke Island),Hanford Res- ervation and thc number of successful nests, 19jj-1981.

Fitzner and Rickard occurredin 1965, *,hen a siogle coyote killed sevengeese, desuoyed eleven oests,and nas probably responsiblefor the desertionof 28 nests. In the fall of 1!66, two tock lesidenceon the island,resultiog in the failure of all but one of 27 nestsduring the 1967 nesting season.In Febtuary 1968,two coyotesrvere shot,and the oumber of goose nestsslightly increasedrhat sa:neyear. Since1970, coyotes have been more or lessperma- nent residentson locke Island, and coyoteremoval was not practicedagain until 1978. By 1980 the Game Departmeot had killed 159 coyotesalong the Hanford Reach (D. Flohr, pers. conm.). Nesting geeseale still absent fron-r Locke Islaod, while coyotes continue to petsist. In fact, a family of coyoteshas taken up residenceon the island despire efforts to control them. Clearly, the survival of this coyote family shon'sthat "general" coyote control is ineffecrive and that the Locke Island situation emphasizes the iriportance of individual aoimalsas predators.Hanson and Eberhardt (1971) noted the sarrrepredator situation in their study and poinred out rhat ao indiscrimioate coyote suppressionptogran conductedduring their scudyprovided no increaseof rhe nesting CanadaGoose population. Locke Islaod has also undergonesome changes in plant com- munity composition io the past ciecadeby grazing of feral cattle. Cheatglass(Btam*t ,ectorLtu) now dominatesnuch of the island formerly dominated by native drylanC perenoialforbs and grasses-These vegetative changes may have had son-reeffect on goose nestiog,b rt the overriding factor has clearly beeo the presenceof coyotesorr the islafld during the nesting season. To better understandthe importanceof Locke Island, one oeed only to examine rhe total number of nestsobserved on the Hanford Reachisland since 19)3. The mean of 135 oestsper year since 1971 is los.er thao 215 nestsper year observedduring 1!!l- 1970 (Hanson and Eberhardt 1970). This decreasecoincides with the near complete termioatioflof nesringon Islandsl,4,6, 8, 10, 13, 14 ( Dewaard1981). Theseislands suppoted 51 perceotof total oestsduriog 1953-1970bur lessthan 4 perceot since 1970. Island 6 supported37 percent of rhe total nestsprior to 1970 but less than 2 percent since then. De\{/aard (1981) analyzedBatelle's data set oo Canada Goose nesting performance,eliminating Islaod 6 in his analysis.He found aonualmears of 136 (S.E.- /) aod 128 (S.E.:5) nestsper yearduring 1953-1970and 1971-1981,respectively. Tlresemeaos were not significantlydifferent (.t - O.792,24 df, P)0.40). Loss of nesting on lsland 6 alone clearly accountsfor the overall drop in number of flestsover the entire study area. Since the CanadaGoose resting population appearsto be fluctuating around 110 nests per year, we cen assumethar rhe Hanford islands are supportiog the ma-\imum number of nesting geese. The averageclutcir size and percent successfulnests have remained neatly unchangedfrom the 1953-i970 to 1971-1981periods, indicating that the locai goosegrpulation is reproductivelyhealthy. An increasein total production of geese,however, is not expectedunless oesting on Isiancl6 is restoredto Leveispreseflt io the early 1960s.

Acknowledgmenls We thank W. C. Hanson for his critical review of the manuscript and N(/. T. Hinds, R. H. Sauer,L. E. Ebethardt,B. D. DeWaard, and J. L. N(arren for their field assistaoce. This study was supportedby the U.S. I)epartmeot of Energy uoder Contract DE-AC06- 76RLO 1810.

(.rnaCaL,r.,use \e.Lin! on the Colunbia. l',-l-l')sl 27r Lileralule Cited

Cushing, C. I., D. G. Watson, A. J. Scoft, ard J. M- GBritisen. 1980. Decline of tadiom:clides in Columbia Rirer b;ora. Pa(ific Norrhwesl LaborarorvPNL-J269. DeVaard. B. K. 1981. Reproduction and Ecology oI Canada ceese on the Hanford Reservation: l95l l')80 M.S. rhesis,Vasbington SrareUniv., Pullman- Fickeisen,-D. H.. D. D. Daubie, D. A.\eitzel. W. H. Rickard, R. L Slaggs. and J. L. Varren. 1980- Aquatic and Riparian Resource Study of rhe Hanford Reach, Columbia"River, \vash- hgron, Report ro.U.S. A_rmy.Corps of Engineers, Seattle District. Battelle, pacific Northwest Laoolatolies. Krchland. wash. Hanson, \7. 8., and L. L. Eberhardt. 1971. A Columbia Caoada Goose population, 1950"1910. Vildlife MonosiaDh 28. lia/. Rickard,. H.,. and,H.-A,. Swgaqy, 1977. Radionuclides in Canada goose eggs, pp. 62j-627, in: Dntcke\ IL, aod R. E. Vilduns (eds.). Biolosical jmplicarioos of m&ik in the envi, ronmenr. Pro-ceedings.oJ the fifteenrh annual Haoford Ljfi Sciences Symposium at Rich- lafld, $74. Sponsored by Baftellq Pncific Northwest Laboratories and Division of Bio, medicxl and Envirorimenral Research and Development Administration. ERDA Symposium series: 42 CONF-750929.

Receiaed.Jauaty 29, 1982 Accepted,fot publ;ca,tionAugutt 10, 1982

272 Fitzner and Rickard