Long-Term Prairie Falcon Population Changes in Relation to Prey Abundance, Weather, Land Uses, and Habitat Conditions

Home , Prairie falcon

University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln

USGS Staff -- Published Research US Geological Survey

2-1999

Karen Steenhof USGS Forest and Rangeland Ecosystem Science Center

Michael Kochert USGS Forest and Rangeland Ecosystem Science Center, [email protected]

Leslie B. Carpenter Raptor Research Center, Boise State University, 970 Lusk Street, Boise, ID

Robert N. Lehman USGS Forest and Rangeland Ecosystem Science Center

Follow this and additional works at: https://digitalcommons.unl.edu/usgsstaffpub

Part of the Earth Sciences Commons

Steenhof, Karen; Kochert, Michael; Carpenter, Leslie B.; and Lehman, Robert N., "Long-Term Prairie Falcon Population Changes in Relation to Prey Abundance, Weather, Land Uses, and Habitat Conditions" (1999). USGS Staff -- Published Research. 44. https://digitalcommons.unl.edu/usgsstaffpub/44

This Article is brought to you for free and open access by the US Geological Survey at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in USGS Staff -- Published Research by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln. The Condor, Vol. 101, No. 1 (Feb., 1999), pp. 28-41 The Condor 101:28-41 ? The Cooper Ornithological Society 1999

LONG-TERMPRAIRIE FALCON POPULATIONCHANGES IN RELATIONTO PREY ABUNDANCE, WEATHER,LAND USES, AND HABITATCONDITIONS1

KAREN STEENHOFAND MICHAELN. KOCHERT USGS Forest and RangelandEcosystem Science Center,Snake River Field Station,970 LuskStreet, Boise, ID 83706, e-mail: [email protected] LESLIEB. CARPENTER RaptorResearch Center,Boise State University,970 LuskStreet, Boise, ID 83706 ROBERTN. LEHMAN USGS Forest and RangelandEcosystem Science Center,Snake River Field Station,970 LuskStreet, Boise, ID 83706

Abstract. We studied a nesting populationof PrairieFalcons (Falco mexicanus)in the Snake River Birds of Prey NationalConservation Area (NCA) from 1974-1997 to identify factorsthat influenceabundance and reproduction.Our samplingperiod included two major droughtsand associatedcrashes in Townsend'sground squirrel (Spermophilus townsendii) populations.The numberof PrairieFalcon pairs found on long-termsurvey segments declined significantlyfrom 1976-1997. Early declines were most severe at the easternend of the NCA, where fires and agriculturehave changednative shrubsteppehabitat. More recent declines occurredin the portionof canyon near the OrchardTraining Area (OTA), where the IdahoArmy National Guard conducts artillery firing and tankmaneuvers. Overall Prairie Falcon reproductiverates were tied closely to annualindexes of groundsquirrel abundance, but precipitationbefore and duringthe breedingseason was relatedinversely to some measures of reproduction.Most reproductiveparameters showed no significanttrends over time, but duringthe 1990s, nesting success and productivitywere lower in the stretchof canyon near the OTA than in adjacentareas. Extensive shrub loss, by itself, did not explain the patternof declines in abundanceand reproduction that we observed.Recent military training activities likely have interactedwith fire and livestock grazingto create less than favorable foraging opportunitiesfor PrairieFalcons in a large part of the NCA. To maintainPrairie Falcon populationsin the NCA, managersshould suppresswildfires, restore native plant communities,and regulatepotentially incompatible land uses. Key words: Falco mexicanus, habitat, land uses, populations,Prairie Falcon, reproduction.

INTRODUCTION Between 1979 and 1997, the NCA experi- Exceptionally high densities of nesting Prairie enced extensive habitat change due to drought, Falcons (Falco mexicanus) in the Snake River wildfires, livestock grazing, and military training Canyon of southwestern Idaho were first rec- activity (Kochert and Pellant 1986, U.S. Dept. ognized in the 1940s (M. Nelson, pers. comm.). Interior 1996). Large areas of native shrub and In 1993, Congress established the Snake River perennial grass communities have been convert- Birds of Prey National Conservation Area ed to homogeneous stands of non-native annual (NCA) to protect 15 species of nesting raptors, vegetation. These changes have affected distri- of which the Prairie Falcon is most numerous bution and abundance of Townsend's ground (U.S. Dept. Interior 1979, 1995). Prairie Falcon squirrels (Spermophilus townsendii; Van Home nesting densities in the NCA are higher than in et al. 1997), the primary prey species of Prairie any other area reported in the literature, and it Falcons in the NCA (Steenhof and Kochert was once estimated that the NCA provided hab- 1988). Densities and reproductive success of itat for up to 5% of the species' nesting pairs many diurnal raptor species are closely related (U.S. Dept. Interior 1979, Steenhof 1998). to the abundance of their prey (Smith and Mur- phy 1979, Smith et al. 1981, Korpimiki 1984). 'Received 15 April 1998. Accepted 12 October Weather, especially precipitation, also affects 1998. nesting success of some birds of prey (Olsen and [28] PRAIRIE FALCON POPULATIONS 29

Olsen 1989, Kostrzewa and Kostrzewa 1990, measurements(Bentley and Hardyman,unpubl. 1991), and weather sometimes interacts with data). We defined four different study stratain prey availabilityto influenceraptor reproduction the NCA canyon (Fig. 1) based upon the poten- (Steenhof et al. 1997). Few investigations,how- tial for military training to affect cliff-nesting ever, have collected adequate data to evaluate falcons: the west-central stratum, a 40-km long-termpopulation changes of predatorsin re- stretchof canyon immediatelysouth of the OTA lation to prey and weatheras well as anthropo- that encompassed radio-tagged pairs found in genic disturbances.The purposeof this paper is the OTA 30% of the time (Marzluff et al. to assess whetherPrairie Falcon abundance,re- 1997); the>- west stratum, an adjacent 20-km production,and distributionin the NCA changed stretch of canyon southwest of the OTA; the between 1974 and 1997, and to identify factors east-centralstratum, a 30-km stretch near C. J. associated with population changes. We were Strike Reservoir,southeast of the OTA; and the particularlyinterested in whetherPrairie Falcon 20-km east stratumnear Hammett(Fig. 1). nesting densities and reproductionwere influ- enced by prey densities,precipitation, vegetation PRAIRIEFALCON ABUNDANCE changes, livestock grazing, or military training We surveyed portions of the canyon to locate activities. nesting pairs during 17 years from 1974-1997. We surveyedthe entireSnake River Canyonand METHODS major side canyons within the NCA for Prairie STUDYAREA Falcon nesting activity in eight years: 1976- 1978 and 1990-1994, and we sampled certain The study areaincluded 2,430 km2of public and stretches of canyon in nine additional years. private lands within the NCA in southwestern From 1979-1984 and in 1997, we surveyed the' Idaho (42050'N, 115050'W). Nesting surveys fo- north side of three segments of the canyon (Fig. cused on a 130-km stretch of the Snake River 1). These "long-term survey segments" con- Canyon between Walters Ferry and Hammett tained 18% of all known nesting areas in the Elevation of the (Fig. 1). canyon ranges from west stratum,61% of nesting areas in the west- m 700 above sea level near the floor to 920 m central stratum,and 38% of nesting areas in the at the canyon rim. Basalt cliffs range in height east stratum. 2 from to 125 m. Topographyabove the canyon Observersrecorded nesting activity of Prairie is generally flat or slightly rolling with a few Falcons and plottednest locations on aerialpho- isolated buttes. Annual precipitation averages tographsand 1:24,000-scale USGS topographic 15-25 cm and occurs mainly from November maps. We defined a historical nesting territory throughApril; summersare hot and dry. Native as any area of cliff where a PrairieFalcon pair vegetation is characteristic of a shrubsteppe was found in one or more years but where no community with big sagebrush (Artemisia tri- more than one PrairieFalcon pair nested in the dentata), shadscale (Atriplex confertifolia), and same year (Newton and Marquiss1982). A nest- winterfat (Krascheninnicovia lanata) associa- ing territorywas consideredoccupied if we ob- tions (U.S. Dept. Interior1996). Livestock graz- served territorialdefense, courtship,or other re- ing occurs throughoutthe area.The IdahoArmy productiveactivity. National Guardconducts trainingmaneuvers in Methodsin the 1970s and 1980s. From 1976- the 56,000-ha OrchardTraining Area (OTA) in 1988, we systematically searched cliffs in the the central part of the NCA (Fig. 1). The Na- Snake River Canyon for Prairie Falcons tional Guarduses the OTA primarilyfor artillery throughout the nesting season. Observers firing, armoredvehicle training,helicopter train- walked along the rim of the canyon or below ing, bivouacking, and small arms firing (U.S. cliffs, stopping periodicallyto search for nests, Dept. Interior1996). and sometimes soliciting responses from terri- We divided the NCA into 10-km stretchesof torial falcons by shouting, clapping hands, or river and side-canyonsto assess the distribution tossing rocks from the cliff top. Incidentalob- of Prairie Falcon pairs. We computed the servationsbegan in Januaryeach year,with sys- amountof cliff area in each stretchfrom studies tematic searches startingin March and continu- that interpretedaerial photographsusing stan- ing throughout the breeding season. We sur- dard parallax methods and field verification veyed most canyon segments at least once dur- 30 KAREN STEENHOFET AL. ing each stage of the nesting cycle (courtship, classified pairs as "laying" if an adultentered a incubation,and brood rearing). cavity and remainedwithin the cavity for > 1 Methods in the 1990s. In 1990, we surveyed hr duringthe incubationperiod. In all years, we the canyon four times, attemptingto duplicate consideredlaying pairs successful if 1 young the 1970s walking surveys. For comparison,we reached 30 days of age (Steenhof 1987).- We also surveyed cliffs from observationpoints on used a photographicaging-key (Moritsch 1983) the canyon floor.Results from walking and point to age young duringnest climbs and remote ob- surveys were similar in 1990, but point surveys servations.Median hatch dates for broods were were more efficient in locating pairs per unit ef- calculatedby backdatingfrom estimatedages of fort (USGS, unpubl.data). From 1991-1997, we young. conducted surveys from observation points in To ensure that productivity comparisons standardized 1-km segments of the canyon. among years were not biased by researcherma- Some side draws were surveyed by walking. nipulations,we excluded nesting attemptswhere Each point was surveyed for 2 hr during each nest site enhancements,manipulations of eggs or monthly survey. When a survey stretch was < young, or disease treatmentoccurred. We also 1 km, survey time was proportionatelyless. excluded informationfrom five nesting attempts Only one side of the canyon was surveyedfrom where investigatorscaused egg or chick mortal- each observationpoint except where the canyon ity. Analyses of percentof pairs successful were was very constricted.We conducted three full based on pairs selected for study prior to each surveys each year from 1991-1994: one in nesting season. The analysis of percentof laying March and early April, one in May, and one in pairs successful included any pairs with nesting June. attempts confirmed during incubation and known outcomes (Steenhof and Kochert 1982). PRAIRIEFALCON REPRODUCTION Brood size at fledging was the mean numberof We assessed Prairie Falcon occupancy, nesting young in nests where 1 young reached 30 success, and productivityfrom a sample of nest- days of age, and was based- on all successful ing territorieseach year from 1974-1983, 1991- pairs for which complete counts were obtained. 1994, and 1997. Before each nesting season, we To estimatenumber fledged per pair and per lay- identifiedat least 20 nesting territoriesknown to ing pair, we multiplied percent success and have been occupied in earlieryears. We tried to brood size at fledging (Steenhof and Kochert locate pairs at these "preselected"nesting ter- 1982). ritoriesduring courtship, ascertain their breeding GROUND ABUNDANCE status during incubation, and observe them as SQUIRREL often as needed to determine occupancy and We used data from live trappinggrids to esti- nesting success. Some nests were entered and mate relative yearly abundanceof Townsend's inspectedfor eggs, young, prey remains,or any ground squirrels in the NCA. For 1975-1982, sign of reproductiveactivity; others only were we used estimated densities of adult squirrels observed from a distance.From 1974-1978, the based on unpublisheddata collected by Smith preselected sample included all territoriesiden- and Johnson (1985) and Johnson et al. (1987). tified by Ogden (1973) within a 72-km stretch Juvenile densities from 1975-1982 were calcu- of the NCA canyon. From 1979-1983, we se- lated using the no recruitmentspecial case of the lected a different set of nesting territoriesfrom Jolly-Sebermodel for open populations,as de- throughoutthe NCA each year, using stratified scribed by Van Home et al. (1997). For 1990- random sampling with proportionalallocation 1994, we used adult and juvenile densities re- according to falcon abundance in canyon portedby Van Home et al. (1997). stretches.We pooled data from radio-taggedand LANDUSES AND HABITAT CONDITIONS control pairs because radios did not influence PrairieFalcon reproduction(Vekasy et al. 1996). We assessed habitatconditions in the predicted In all years we confirmedlaying if an occu- foraging areas of falcons from each of the four pied nesting territory contained an incubating canyon strata (Fig. 1) on a Geographic Infor- adult, eggs, young, or any other indicationthat mation System using data collected by other in- eggs were laid (e.g., fresh eggshell fragmentsin vestigators.Predicted foraging areas were based fresh nesting material). In the 1990s, we also on home ranges determinedby radio telemetry PRAIRIEFALCON POPULATIONS 31

WALTERS FERRY *BoiseNational AreaConservation

...... ::...... ::::;ii:. ?: .x4:M i

Orchrd rainng rea RESERVOIR?CJSTRIK? HAMET ForagingAreas: X\,'")'?:?'iI I::'. : W est?'?i'ii? Foraging Areas West-entrLog-tem Surey I Sgment EAS 1 Ddication Pont Eastcentral to Swan alls km ...... 2 all Poit toChafin ill](40km) (4 0 5 10 ilom ter East Waterfall 3 Draw to Pump Draw (10 km). West West-central Long-term Survey Segments EAST I .East-central 1 Dedication Point to Swan Falls (4 km) 2 Balls Point to Chattin HilI (40 km) 0 5 Kilometers East 3 Waterfall Draw to Pump Draw (10 km)

FIGURE1. Locationof the OrchardTraining Area, standardized survey segments,canyon strata, and associated foragingareas in and near the Snake River Birds of Prey NationalConservation Area.

(Dunstan et al. 1978, Marzluff et al. 1997). Prai- as present or absent on four segments of each rie Falcons in the NCA foragedover large (mean transect.The sum of the scores gave each tran- = 300 km2), undefendedareas extending up to sect a grazing intensity score of 0 (absent), 1 22 km north of the canyon (Marzluff et al. (low), 2 (medium low), 3 (medium high), or 4 1997). (high). We tabulatedthe percent of transectsin We assessed habitat change in each stratum each stratumthat had signs of livestock use and by comparinga 1979 vegetationmap, developed calculated the mean grazing intensity score for through visual interpretationand field verifica- transects within each stratum that had signs of tion of low level aerial photographs(U.S. Dept. livestock use. We categorized Townsend's Interior1979), to a 1994 vegetationmap, which ground squirrelhabitat potential as low or high was based on supervised classificationof Land- on transectswithin each stratum,based on the sat thematicmapper satellite imagery (Knick et amount of native perennialgrass cover (< or > al. 1997). We assessed shrub loss from 1994 to 5%, respectively; U.S. Dept. Interior 1996) on 1997 from Bureau of Land Management and 589 sites sampledby S. T. Knick and S. E. Watts Idaho Army National Guardmaps that depicted (unpubl.data) from 1991-1994. boundaries of fires that occurred from 1994- 1996. STATISTICALANALYSES We assessed livestock grazing intensityin the We used SYSTAT(SPSS Inc. 1997) and STAT- same four strata from S. T. Knick and S. E. XACT (Cytel Software 1989) for statisticalan- Watts' unpublishedcounts of cattle and sheep alyses. Significancewas evaluatedat P = 0.05. fecal material on 485 randomly selected tran- We used precipitationdata from the Boise Air- sects (USGS, unpubl. data). From 1991-1994, port weatherstation for all weatheranalyses. We observersclassified fecal matterof each species used Kruskal-Wallistests to compareinitial nest- 32 KAREN STEENHOFET AL.

220

210 205 206

200 194

190 187 0 182 183 0) 180 .E Z 170 160 160

150 I , 76 . 77 .. 78 90 .I91 92 93 94 Year FIGURE2. Numberof PrairieFalcon pairs in the SnakeRiver Birds of PreyNational Conservation Area in the eightyears that full surveyswere conducted. ing densities by strata.We used Mann-Whitney times nested within 50 m of each other.Numbers tests to assess differences in falcon abundance were lowest in the east stratum(0.8 pairs km-') in the 1970s and 1980s versus the 1990s. We and intermediatein the east-central(1.29 pairs assessed trends throughtime by runningsimple km-') and west-central(1.61 pairs km-') strata. correlationsbetween year and selected variables. From 1976 to 1978, distances between nearest We used two-factor (year and strata)log-linear adjacentpairs averaged646 m. models to test for differencesin occupancy and The amount of cliff surface area explained success rates.We used two-factor(year and stra- 91% of variationin numberof falcon pairs per ta) analyses of varianceto assess differences in 10-km survey stretch in 1976 (P < 0.001). brood size at fledging with hatch date as a co- Number of pairs km-2 of cliff also varied sig- variate. Relationships between ground squirrel nificantly by strata (Kruskal-WallisH3 = 10.5, abundance,precipitation, and falcon reproduc- P = 0.02) in the 1970s. When assessed in relation and abundance were assessed by simple tion to amount of cliff area available, nesting correlationand multipleregression analyses. We densities were highest in the west-centralstra- used log-likelihood G-tests to compare propor- tum (mean = 45.8 pairs km-2), nearly as high tions of transectswith sheep and cattle use by in the west stratum(39.7 pairs km-2), and much strata.We transformedcattle and sheep grazing lower in the east (35.4 pairs km-2) and east-cen- intensityscores using a square-rootfunction and tral (34.7 pairs km-2) strata. comparedscores among stratausing analysis of variance with Bonferroni adjusted paired con- CHANGESIN PRAIRIEFALCON ABUNDANCE trasts. Total numberof PrairieFalcon pairs nesting in the NCA from 160 to 206 in the RESULTS ranged eight years that full surveys were conducted(Fig. 2). PRAIRIEFALCON DISTRIBUTION IN THE1970S Mean numberof pairs in the early survey years Prairie Falcons were not evenly distributed (1976-1978) did not differ from the mean num- throughout the NCA when surveys began in ber in 1990-1994 (Mann-WhitneyU = 11.0, P 1976. From 1976-1978, numberof pairs per ki- = 0.30); althoughnot significant,correlation an- lometer of river varied significantly (Kruskal- alyses suggestedan overall decline in numberof Wallis H3 = 8.3, P = 0.04) among strata.Num- pairs through time (r = -0.64, n = 8, P = 0.09). ber of pairs per kilometer was highest in the Numberof pairs was highest duringthe firsttwo west stratum(mean = 3.28 pairs km-') where years of complete surveys, but numbersdeclined cliffs are highest and where falcon pairs some- in 1978. Numberof pairsfound in 1990 was just PRAIRIE FALCON POPULATIONS 33

100 315.2, P < 0.001). Occupancy rates declined -.-All90 Segments 90 West-central significantly during the 1990s (r = -0.96, n = 80 --- West 5, P = 0.01); were in 1992 (68%) 700 East they highest _ and lowest in 1997 when pairs occupied only IL 60 o50 37% of traditionalterritories. Mean occupancy rates during the 1990s were highest in the west i 40 and east-central strataand lowest E 30 (73%) (67%) in the east (47%) and west-central strata. Z 20 (51%) Occupancyrates in the west-centralstratum declined from 1992 to 1997, whereas rates in the 77 79 81 83 85 90 92 94 96 east stratumwere above average in 1997 (64%). Year PRAIRIEFALCON REPRODUCTION FIGURE3. Numberof PrairieFalcon pairs in long- term survey segmentswithin the Snake River Birds of From 1974-1997, an averageof 63% of all Prai- Prey NationalConservation Area, 1976-1997. rie Falcon pairs successfully raised young, and an average of 70% of laying pairs were successful (Table 1). Success rates showed no sig- slightly higher than that found in 1978. In 1994, nificant trends over time (r = -0.205, n = 15, we identifiedonly 160 PrairieFalcon pairsin the P = 0.46 for all pairs; r = 0.007, n = 22, P = NCA, fewer than in any year that complete sur- 0.98 for laying pairs). Successful pairs fledged veys of the area were conducted. Declines oc- an average of 3.9 young for all years; mean curred in many of the 10-km stretches we sur- brood size at fledging increasedslightly but sig- veyed. By 1994, 7 of 19 10-km stretches had nificantly from 1974-1997 (r = 0.46, n = 22, P Prairie Falcon densities below levels recorded = 0.03). From 1991-1997, mean brood size at from and 12 stretches were un- 1976-1978, fledging (4.19) was significantlyhigher (t2.7 - changedfrom 1976-1978 levels. No stretchesof -4.64, P < 0.001) than from 1974-1983 (3.76). canyon had more pairs than observed from Overall productivity averaged 2.46 young per 1976-1978. pair from 1974-1997 (Table 1). The medianpro- Numberof pairs in the threelong-term survey ductivity rate over the same time period (2.42) segments correlated closely with number of exceeded the median replacementstandard of 2 pairs in the entire study area in years when full young per pair thatRunde (1987) calculatedwas surveys were conducted (r = 0.93, n = 8, P = necessary to maintaina PrairieFalcon popula- 0.001). From 1976-1997, numberof falcon pairs tion through time. Number of young per pair nesting in the three long-term survey segments showed no significant trends over time (r = declined significantly (r = -0.66, n = 15, P = -0.02, n = 15, P = 0.93). 0.007). In 1997, the 37 pairs in these stretches representeda 46% decline from the 1976 level PRAIRIEFALCON-GROUND SQUIRREL RELATIONSHIPS of 68, and a 51% decline from the peak count of 75 recordedin 1984 (Fig. 3). The decline in Ground squirrel numbers were relatively high number of pairs on the three long-term survey when studies began in 1975 and 1976 (Fig. 4). segments occurred mainly in the west-central A winterdrought prior to the 1977 breedingsea- segment (r = -0.65, n = 15, P = 0.009). Pop- son caused squirrelsto suspend reproductionin ulations in the west survey segment showed no 1977 and enter torpor 4-6 weeks earlier than significant change (r = -0.25, n = 15, P = usual (Smith and Johnson 1985). Densities were 0.37). A small, but long-term decline was sig- low in 1978 due to the absence of a yearling nificant in the east segment, even though num- cohort. Numbers recovered slowly from 1979- bers in the east increasedfrom 1994 to 1997 (r 1982 (Fig. 4). In 1991, ground squirrel abun- = -0.66, n = 15, P = 0.008). dance was higher than in the 1970s, and it in- During the 1990s, Prairie Falcons occupied creased to a peak level of 24.5 squirrelsha-' in 37-68% of the historical territorieswe sampled 1992. A spring 1992 droughtfollowed by a se- each year (mean ? SD = 57 + 12%, n = 5). vere winter in 1992-1993 affected overwinter Occupancyrates differed significantlywith both survival and drasticallyreduced ground squirrel year (X24 = 55.8, P < 0.001) and strata(X23 = abundancein 1993 and 1994 (Van Horne et al. 34 KAREN STEENHOF ET AL.

TABLE 1. Nesting success and productivityof PrairieFalcons in the Snake River Birds of Prey National ConservationArea, 1974-1997. Sample sizes are in parentheses.

% of pairs % of laying pairs Brood size at No. fledged/ No. fledged/ Total Year successful successful fledging laying paira pairb fledgedc 1974 69 (16) 64 (11) 3.94 (17) 2.52 2.72 1975 76 (25) 86 (22) 3.76 (29) 3.23 2.86 1976 75 (24) 81 (31) 3.98 (43) 3.22 2.99 613 1977 55 (20) 67 (33) 3.64 (33) 2.44 2.00 412 1978 50 (28) 18 (35) 3.67 (30) 0.66 1.84 335 1979 65 (43) 75 (32) 3.73 (30) 2.80 2.42 1980 56 (52) 65 (54) 4.03 (36) 2.62 2.26 1981 80 (51) 83 (46) 4.05 (22) 3.36 3.24 1982 41 (29) 69 (26) 3.53 (17) 2.44 1.45 1983 71 (31) 73 (26) 3.24 (17) 2.37 2.30 1984 -- 88 (16) 3.95 (19) 3.48 1985 -- 93 (14) 4.08 (13) 3.79 1986 -- 44 (18) 3.43 (7) 1.51 1987 -- 70 (37) 4.32 (31) 3.02 1988 -- 75 (8) 3.86 (7) 2.90 1989 -- 83 (6) 3.67 (6) 3.05 1990 -- 68 (41) 4.06 (33) 2.76 1991 78 (68) 89 (37) 4.08 (49) 3.63 3.18 582 1992 79 (85) 86 (64) 4.23 (53) 3.64 3.34 648 1993 37 (91) 47 (53) 4.14 (37) 1.95 1.53 286 1994 61 (83) 69 (58) 4.34 (44) 2.99 2.65 424 1997 51 (37) 56 (27) 4.14 (14) 2.30 2.11 X - SD all years 63 ? 14 70 ? 17 3.90 ? 0.29 2.76 ? 0.74 2.46 ? 0.61 a Calculated as the product of % of laying pairs successful and brood size at fledging. b Calculated as the product of % of pairs successful and brood size at fledging. Data are not available for 1984-1990 because Prairie Falcon pairs were not in those c preselected years. Data available only for years with complete counts of pairs for the entire area.

1997). In 1997, transect sampling in portions of NCA correlated strongly (Ps < 0.01) with both the NCA, coupled with subjective observations, juvenile and overall ground squirrel density. To- suggested that overall squirrel populations had tal number of young fledged from the long-term increased from 1994 levels. survey segments, which was based on four ad- Most measures of Prairie Falcon reproduction ditional years, also correlated significantly and correlated positively with estimated ground positively with juvenile and overall squirrel den- squirrel abundance (Table 2). From 1976-1994, sities (Ps < 0.05). We also found significant cor- total number of Prairie Falcons fledged from the relations (Ps < 0.05) between both percent of falcon pairs successful and number of falcons and and overall 30 fledged per pair juvenile squirrel P Juveniles densities. Number fledged per laying pair cor- - Adults related significantly only with overall squirrel _25 densities, and percent of laying pairs successful 200 correlated significantly only with adult squirrel densities. Three variables (brood size at fledg- 3 15 ing, total pairs, and pairs in the long-term survey o0 segments) did not correlate significantly with any of the measures of ground squirrel abun- 5 dance, but number of pairs was more closely related to adult squirrel abundance (P = 0.08) than to juvenile or overall squirrel abundance 76 78 80 82 84 90 92 94 (Ps > 0.38). We found no relationship between Year number of nesting pairs and ground squirrel FIGURE 4. Estimated densities of Townsend's abundance in the previous year (r = 0.56, n = groundsquirrels (n ha-') in the NCA, 1975-1994. 6, P = 0.25). PRAIRIEFALCON POPULATIONS 35

TABLE 2. Correlationcoefficients between estimatedTownsend's ground squirrel abundance and measuresof PrairieFalcon reproductionin the Snake River Birds of Prey NationalConservation Area, 1975-1994.

Groundsquirrel density PrairieFalcons n Adult Juvenile Total Numberof nesting pairs 7 0.70 0.28 0.39 Numberof pairs in the long-term survey segments 11 0.52 0.23 0.32 Percentof pairs successful 12 0.44 0.65* 0.64* Percentof laying pairs successful 12 0.57* 0.51 0.56 Numberfledged per laying pair 12 0.50 0.55 0.57* Brood size at fledging 12 -0.20 0.25 0.14 Numberfledged per pair 12 0.37 0.66* 0.63* Total numberfledged 7 0.75* 0.88** 0.89** Numberfledged in the long-term survey segments 11 0.49 0.63* 0.64* * P < 0.05, ** P < 0.01.

Two measures of Prairie Falcon reproduction of the variation in these measures when we con- that did not correlate strongly with ground squir- sidered squirrel abundance and weather together rel numbers were related inversely to the amount in multiple regressions (Ps > 0.10). of precipitation prior to and during the onset of the breeding season. Brood size at fledging and SPATIALVARIATION IN FALCON REPRODUCTION number fledged per laying pair correlated inversely with total precipitation from November During the 1990s, nesting success of Prairie Fal- to April (r = -0.49 and -0.51, respectively, n con pairs varied significantly among study strata = 22, P = 0.02) and also from February through (X23= 97.8, P < 0.001) and years (X24 = 30.0, April (r = -0.52 and -0.43, respectively, P = P < 0.001). After the 1992 drought, nesting suc- 0.01 and 0.05, respectively), but not with pre- cess dropped sharply in the west-central stratum, cipitation from April through June (Ps > 0.56). much smaller declines occurred in the west and None of the other measures of falcon reproduc- east-central strata, and success rates in the east tion was related to precipitation (Ps > 0.06), and showed no change (Fig. 5). In 1997, the success precipitation did not explain a significant portion rate in the west stratum rebounded from 1994 levels, but the rate in the west-central stratum declined. Until we did not have 100 1991, adequate sample sizes to analyze success by strata by 90- year. However, when we pooled data within our 80- three samplingperiods (1974-1983, 1991-1994, and 1997), we found that success rates in the west-central stratum declined significantly with time (G2= 10.2, P = 0.006; Fig. 6); rates in the east and west strata were unchanged (G2 = so C) 0.006, P = 0.99 and G2= 0.97, P = 0.62, re- 40 spectively). Mean brood size at did not 30- fledging vary sig- West strata the 1990s = SWest-central nificantly by during (F,385 20 East-central = -- - East 1.0, P 0.38). Hatch date was a significant co- 10 ...... variate in the analysis of brood size at fledging = 91 92 93 94 96 96 97 (F1,182 21.7, P < 0.001) with smaller broods Year being more common late in the nesting season. Number of young fledged per pair fell below the FIGURE 5. Annual Prairie Falcon nesting success rates in four strata within the Snake River Birds of minimum replacement standard of 2 (Runde Prey National Conservation Area, 1991-1997. The 1987) in 1993 and 1997 in the west-central stra- east-centralstratum was not sampledin 1997. tum, and in 1993 in the east-central stratum, but 36 KAREN STEENHOF ET AL.

80 4% in the west-central, 18% in the east-central,

70 and 22% in the west strata. By 1979, the pro- portion of the east stratum in shrubs was ap- 0 60 proximately half that of the other strata (Fig. 7). Between 1979 and 1994, the nature and dis- V) so50 tribution of habitat change differed from that before 1979. Less than 2% of the native rangeland within Prairie Falcon foraging ranges was con- 40t verted to agriculture between 1979 and 1994. 30 Fires, on the other hand, caused extensive shrub West West-central East-central East loss, primarily between 1981 and 1985 (Kochert FIGURE6. Percentof PrairieFalcon pairssuccessful and Pellant 1986). Most shrub loss between in four study strataand threetime in the Snake periods 1979 and 1994 occurred in the east-central and River Birds of Prey NationalConservation Area. Open bars represent1974-1983, black bars represent1991- west-central strata (Fig. 7). By 1994, fires and 1994, and hatchedbars represent1997. The east-cen- agriculture had left only 4-35% of the habitat in tral stratumwas not in 1997. sampled each stratum in shrubs (Fig 7). From 1994- 1996, fires consumed an additional 38,700 ha, in the west stratum. Most fires that overall differences were not significant among primarily occurred from 1994-1996 in the west-central and strata from 1991-1997 (F3,14 = 1.0, P = 0.42). east-central strata were in previously burned ar- HABITATCONDITIONS AND LAND USES eas, so there was little loss of additional shrub- Habitat alteration occurred in all strata and in all lands. By 1997, native shrub habitat comprised years of our study, but its intensity, distribution, only 4-28% of the area in each stratum. The and timing varied (Fig. 7). By 1979, fires and west-central stratum still had the highest percent agriculture already had fragmented habitat in the shrub cover, even though it sustained extensive east stratum to a much greater extent than in the shrub losses between 1979 and 1994 (Fig 7). other strata. Thirty-five percent of the east stra- Ground squirrel habitat potential, as measured tum was grassland in 1979, compared to only by the amount of native perennial grass cover in 15-18% of the other strata. In addition, 27% of 1991-1994, differed significantly by strata (G3 the east stratum was farmed, compared to only = 70.8, P < 0.001). More transects had high

WEST WEST-CENTRAL EAST-CENTRAL EAST

1979

1994

1997

Shrub OMAgiclneus .Grassland FIGURE7. Percentof PrairieFalcon foragingareas composed of shrub,grass, agriculturaland otherhabitats in four strata,1979-1997. PRAIRIE FALCON POPULATIONS 37 ground squirrel habitat potential in the west that were successful, numberof young fledged (66%) and west-central (58%) strata than the per pair, and total number of young fledged. east (20%) and east-central(26%) strata(G1 = Prairie falcon reproductionwas best predicted 67.3, P < 0.001). by juvenile ground squirrel densities or by a Military training occurred only in the west- combination of juvenile and adult densities. central stratum. Livestock grazing occurred Number of falcon nesting pairs, on the other throughoutthe area, but the proportionof sites hand, was not relatedto juvenile groundsquirrel with signs of livestock use varied significantly abundance and only weakly related to abun- among strata for both cattle (G3 = 10.1, P = dance of adult squirrels.Galushin (1974) pre- 0.02) and sheep (G3 = 19.2, P < 0.001). The dicted that if migratoryraptors find scarce food east stratumhad significantlymore sites used by when they returnto previousnesting areas,they cattle and significantly fewer sites used by may leave to find betternesting conditions else- sheep. The proportionof sites with cattle (G2 = where. P and = P = 1.3, = 0.52) sheep (G2 1.7, 0.42) WEATHER was similar in the three westernmost strata. Extreme affected PrairieFalcons twice Grazingintensity scores at used sites varied sig- droughts during the 24-year study period. Prairie falcon nificantlyamong stratafor cattle (F3,423 = 3.4, P success and declined as = 0.02), but not for = P = nesting productivity sheep (F3,102 1.8, declined aftera win- 0.15). Cattle use was significantlyhigher in the groundsquirrel populations ter in 1976-1977 and Johnson east than in the west stratum(pairwise compar- drought (Smith and after a in 1992 ison with Bonferroniadjustment, P = 0.01); cat- 1985) spring drought (Van tle use in the west-centraland east-centralstrata Home et al. 1997). In contrastto other raptors was intermediateand did not differ (Olsen and Olsen 1989, Kostrzewaand Kostrze- significantly wa rates of Prai- from the other strata(Ps > 0.21). 1990), long-termreproductive rie Falcons were not related to spring precipi- DISCUSSION tationduring the brood-rearingperiod. However, brood sizes at fledging were relatedinversely to PREYABUNDANCE the amountof precipitationprior to and during Our long-termdata on yearly variationin nest- the onset of the breeding season. Successful ing success and productivitysupport earlier con- Prairie Falcons produced large broods in dry clusions (U.S. Dept. Interior 1979) that Prairie years and smaller broods in wet years. Precipi- tation Falcon reproductionin the NCA is closely tied priorto the nesting season usuallyincreas- es the amount of and to groundsquirrel abundance. Overall falcon re- grass weed cover in the NCA and interferewith the of Prai- production in the NCA showed no significant may ability rie Falcons to find and secure Bechard trends through time. Instead it varied among prey. found that cover limited the years, as groundsquirrel populations responded (1982) vegetative of to Swainson's Hawks to two significant droughts. Ground squirrels accessibility prey (Bu- have been and continue to be the most teo swainsoni). Decreased foraging efficiency single associated with tall cover late in the important for PrairieFalcons in the vegetative prey species season could affect Prairie Falcons NCA (Steenhof and Kochert 1988, nesting by Holthuijzen the number of that successful 1990, Marzluff et al. Steenhof and Ko- lowering young 1997). can winter chert (1988) demonstratedthat Prairie Falcons pairs support. Precipitationduring and early spring also could enhance conditions in the NCA were dietary specialists.Prairie Fal- for nest parasitesand thereforedecrease nestling cons showed strong preferencesfor Townsend's survival; McFadzen and Marzluff (1996) re- ground squirrelseven when squirrelswere rare, ported more nestling mortality due to infesta- and falcons had no single importantalternate tions of in the NCA a inves- hematophagousectoparasites (Haema- prey species during 7-year tosiphoninodurus) in PrairieFalcon nests within and Kochert Ground tigation (Steenhof 1988). the NCA during the wet spring of 1993 than squirrelscontain a high amountof fat and pro- during the dry spring of 1992. vide more kilocaloriesper gram than alternative prey species (U.S. Dept. Interior1979). Ground HABITATCONDITIONS AND LAND USE squirrel abundanceduring the breeding season Long-term Prairie Falcon population declines appearedto influencethe percentof falcon pairs may be related to extensive shrub loss that has 38 KAREN STEENHOFET AL. occurred in the NCA. Although Townsend's with fire and livestock grazingto createless than ground squirrelabundance is closely tied to the favorable foraging opportunitiesfor falcons in amount of native perennialgrass cover, ground the west-centralstratum. Radio telemetrystudies squirrelsalso rely on shrubs(U.S. Dept. Interior in the 1990s (Marzluffet al. 1997) showed that 1996). Adult groundsquirrels ate new growthof PrairieFalcons from the west-centralarea were sagebrush,and juveniles ate primarilywinterfat less effective at obtaining Townsend's ground during drought years (Van Home et al. 1998). squirrels than falcons nesting in the west seg- Groundsquirrels in burnedareas dominatedby ment, particularlyduring drought years. West- non-native annual grasses are more vulnerable centralfalcons flew over largerareas, spent less to fluctuationsin food biomass and diversity; time at their nests, and delivered fewer ground populationsin annualgrasslands devoid of shrub squirrels to their nests than falcons from the cover show wider annualfluctuations (Yensen et west stratum(U.S. Dept. Interior1996). al. 1992, Van Home et al. 1997). Optimal Military training could affect Prairie Falcon ground squirrelhabitat likely consists of native foraging efficiency in at least two ways. First, perennial grasslands interspersed with shrub trainingactivity could directly disturb foraging cover (U.S. Dept. Interior1996). falcons and prevent them from securing ade- Early declines in number of Prairie Falcon quate prey. Military training activity did have pairs nesting at the easternend of the NCA may direct, short-terminfluences on Prairie Falcon have been associatedwith habitatfragmentation foraging behavior,which were apparentprimar- caused by fires and agriculturaldevelopment ily on days with intense military training(U.S. that occurred before our study began. Habitat Dept. Interior1996). These periods of intensive patternswe measured,however, did not explain training were relatively infrequent and should recent declines in both abundanceand reproduc- not have affected the ability of PrairieFalcons tion of PrairieFalcons in the west-centralstra- to secure prey over the entire nesting season. tum. Groundsquirrel habitat potential, as mea- However, disruption of foraging at a critical sured by the amount of native perennial grass point in the nesting season could have resulted cover in 1991-1994, was above average in the in nesting failures.Intensive militarytraining in west-centralarea, and indexes of livestock graz- June and July also might interferewith the for- ing levels were similar in the west-centraland aging efficiency and subsequent survival of adjacent strata.At the beginning of our study, post-fledgingPrairie Falcons, but additionalre- the west-centralstratum contained a higher per- search would be needed to determine whether centage of shrubs than the other strata. The this occurs. In Wyoming, PrairieFalcons toler- west-central stratum apparentlyprovided ade- ated disturbancesassociated with low levels of quate foraging opportunitiesfor PrairieFalcons oil developmenton their foragingareas (Squires in the 1970s because density of falcon pairs per et al. 1993). area of cliff was higher than in the other strata, Second, training activity could cause subtle and nesting success was similar to that in other habitatchanges, which we were unable to mea- strata during the 1970s. The west-centralarea sure, that might be detrimentalto ground squir- experiencedextensive shrub loss between 1979 rel populations. The best indirect measure of and 1997, but the amountof loss was similarto habitat suitability, amount of native perennial thatin otherstrata, and by 1997, the west-central grass cover, suggests that ground squirrels stratumstill had a higher percentageof shrubs should be at above average levels in the west- than the other strata. central stratum.Unfortunately, we do not have The single obvious factor distinguishingthe reliable landscape-level data on the relative west-centralstratum is the presence of military abundanceof groundsquirrels in the four strata training.There are no quantitativedata on his- to confirm this. Long-termuse of armoredve- torical levels of military training,but the scope hicles and short-termarmored-vehicle tracking and nature of military training in the OTA experimentsin shrub and grass habitats in the changed in the early 1990s when the National OTA had no detectableeffects on groundsquir- Guardbegan upgradingexisting trainingfacili- rel populationdynamics (Van Home and Sharpe ties and constructingadditional ones (U.S. Bu- 1998). However, tracked vehicles used during reau of Land Management1988). Recent mili- military trainingcause grounddisturbances that tary training activities likely have interacted result in habitat fragmentationand increased PRAIRIE FALCON POPULATIONS 39

dominance by exotic annual plant communities mechanisms that might be involved. Although (Knick and Rotenberry 1997). In addition, areas we have been unable to isolate a single causal like the OTA's artillery impact area, which are mechanism for more pronounced population de- subjected to repeated burning, are less likely to clines in the west-central stratum, it appears that experience natural shrub regeneration than areas the west-central stratum is the area most in need that have burned only once (U.S. Dept. Interior of rest and restoration. Continued and expanded 1996). These altered habitats are not generally monitoring of Prairie Falcon populations and beneficial for Townsend's ground squirrels over their prey throughout the entire NCA will be the long-term (Yensen et al. 1992, Van Home et helpful in distinguishing influences of individual al. 1997). land uses and habitat perturbations. However, The west-central area also may be inherently because land uses and habitat perturbations act less resilient to all forms of disturbance and less both synergistically and cumulatively, it is un- responsive to restoration efforts that land man- likely that we will ever find a single cause for agement agencies have tested so far. In 1979, declines in Prairie Falcon abundance and pro- shadscale communities, which support lower ductivity. A continuing decline in number of densities of ground squirrels than other habitats nesting pairs and a possibly permanent reduction (U.S. Dept. Interior 1979, Smith and Johnson in the area's carrying capacity for Prairie Fal- 1985), dominated large areas of the west-central cons would be inconsistent with goals of the leg- stratum. By 1997, losses of sagebrush in the islation that established the NCA. Managers of OTA, in combination with poor quality shad- the NCA should aggressively suppress wildfires, scale habitats south of the training area, may actively restore native shrubs and perennial have left the west-central stratum with too few grasses, and regulate existing land uses within ground squirrels to support Prairie Falcon pop- Prairie Falcon foraging areas. Subsequent mon- ulations at levels we recorded in the 1970s. itoring, integrated at all three trophic levels, will Timing of habitat alteration also may be im- be necessary to determine whether land use re- portant. Although the total amount of shrub loss strictions and habitat recovery/restoration can in the west-central stratum has been less than in reverse population declines. Management ac- the other strata, it occurred during a narrower tions should be coordinated closely with moni- window of time (1979-1994 versus pre-1979- toring efforts in an adaptive management ap- 1994 for the east and east-central and 1979- proach (Lancia et al. 1996) to determine whether 1997 for the west). The consequences of habitat restrictions and restoration efforts are achieving loss may be more serious when it occurs rapidly. desired results. It also is possible that Prairie Falcons, like Gold- ACKNOWLEDGMENTS en Eagles (Aquila chrysaetos; Kochert et al., unpubl. data), respond to habitat alterations several This paperis a contributionfrom the SnakeRiver Field years after they occur. Extensive fires in the west Station, Forest and Rangeland Ecosystem Science stratum occurred more Center,Biological ResourcesDivision, U.S. Geologi- recently (1995-1996) cal This as of an than in the west-central stratum. If west-central Survey. study began part investiga- tion that led to establishingthe Snake River Birds of falcons are just now responding to habitat Prey NCA; more recently it became part of the U.S. changes that occurred between 1979 and 1994, Bureauof LandManagement and IdahoArmy Nation- al Guard research We are we might similar in the west cooperative project. indebted expect changes to more than 100 and technicians stratum in future biologists who years. helped to collect data in the field. Data collection has Managers of the Snake River Birds of Prey been coordinatedby three differentagencies: first and NCA are facing serious challenges in their at- primarilythe Bureau of Land Management,then the tempts to regulate land uses to protect and en- NationalBiological Service, and most recentlythe Bi- hance face dif- ological Resources Division of the U.S. Geological raptor populations. They equally A. R. Bammannand J. H. Doremus a ficult as to cumulative Survey. played challenges they try repair key role in collecting data in the early years of the damages caused by wildfire, drought, livestock study. We thank J. M. Marzluff,M. S. Vekasy, L. S. grazing, and military training activity. Our data Schueck, and all GreenfalkConsultant personnel for suggest that foraging areas in the west-central their cooperationin the 1990s. We are gratefulfor the use data stratum have been most opportunityto unpublished on prey and hab- seriously compromised. itat collected by severalinvestigators: G. W. Smith,N. If military training activity is affecting Prairie C. Nydegger,D. L. Quinney for the 1979 vegetation Falcons adversely, we do not yet understand the map; N. C. Nydegger, D. L. Johnson, and B. Van 40 KAREN STEENHOFET AL.

Home for data on ground squirrels;and S. T Knick teo buteo), goshawk (Accipiter gentilis) and kes- and S. E. Watts for data on vegetation and livestock trel (Falco tinnunculus). Ibis 132:550-559. use in the 1990s. Special thanks go to G. R. Griggs, KOSTRZEWA,R., AND A. KOSTRZEWA.1991. Winter W. D. Dyer, and E. E. Black for logistical supportand weather,spring and summerdensity, and subse- coordination,to the U.S. Fish and Wildlife Service, quent breeding success of Eurasian Kestrels, Snake River Basin Office for the loan of a jet boat, to CommonBuzzards, and NorthernGoshawks. Auk G. S. Olson who analyzed juvenile ground squirrel 108:342-347. data, to T J. Zarriellofor GIS analyses, and to D. M. LANCIA, R. A., C. E. BRAUN,M. W. COLLOPY,R. D. Parrishfor data automation.We are indebtedto S. T. DUESER,J. G. KIE, C. J. MARTINKA,J. D. NICHOLS, Knick, J. T Rotenberry,and B. Van Home for helpful T. D. NUDDS,W. R. PORATH,AND N. G. TILGHMAN. advice and discussionsand for reviewinga draftof this 1996. ARM! For the future: adaptive resource manuscript.J. Squires and an anonymous reviewer management in the wildlife profession. Wildl. also providedcomments that improved the manuscript. Soc. Bull. 24:436-442. Fundingfor this study was providedprimarily by the MARZLUFF,J. M., B. A. KIMSEY,L. S. SCHUECK,M. E. Bureauof LandManagement from 1974-1989 and by MCFADZEN, M. S. VEKASY, AND J. C. BEDNARZ. Idaho NationalGuard W. S. the Army through Seegar, 1997. The influence of habitat,prey abundance, U.S. ChemicalResearch and Center Army Engineering sex, and breedingsuccess on the rangingbehavior from 1990-1994. The NationalBiological Service and of PrairieFalcons. Condor99:567-584. the ResourcesDivision of the U.S. Biological Geolog- MCFADZEN,M., ANDJ. M. MARZLUFF.1996. Mortality ical Survey provided additionalfunding and support of PrairieFalcons the from 1994-1997. during fledglingdependence period.Condor 98:791-800. MORITSCH,M. Q. 1983. Photographicguide for aging LITERATURECITED nestlingPrairie Falcons. U.S. Dept. Int.,Bur. Land BECHARD,M. J. 1982. Effect of vegetative cover on Manage.,Boise, ID. foragingsite selection by Swainson'sHawk. Con- NEWTON,I., ANDM. MARQUISS.1982. Fidelity to breed- dor 84:153-159. ing area and mate in Sparrowhawks.J. Anim. CYTEL SOFTWARECORPORATION. 1989. StatXactstatis- Ecol. 54:241-253. tical software for exact nonparametricinference. OGDEN,V. T. 1973. Nesting density and reproductive Cytel SoftwareCorporation, Cambridge, MA. success of the PrairieFalcon Falco mexicanusin DUNSTAN,T. C., J. H. HARPER,AND K. B. PHIPPS.1978. southwesternIdaho. M.Sc. thesis, Univ. Idaho, Habitatuse and huntingstrategies of PrairieFal- Moscow, ID. cons, Red-tailedHawks, and Golden Eagles. U.S. OLSEN,P. D., AND J. OLSEN. 1989. Breeding of the Dept. Int., Bur. Land Manage., Denver Federal PeregrineFalcon Falco peregrinus:III. weather, Center,Denver, CO. nest quality and breedingsuccess. Emu 89:6-14. GALUSHIN, V. M. 1974. Synchronousfluctuations in RUNDE,D. E. 1987. Populationdynamics, habitat use populationsof some raptorsand their prey. Ibis and movementpatterns of the PrairieFalcon (Fal- 116:127-134. co mexicanus). Ph.D. diss., Univ. Wyoming, Lar- JOHNSON,D. R., N. C. NYDEGGER,AND G. W. SMITH. amie, WY. 1987. Comparisonof movement-baseddensity es- SMITH,G. W., ANDD. R. JOHNSON.1985. Demography timates for Townsendground squirrelsin south- of a Townsend ground squirrel population in westernIdaho. J. Mammal.68:689-691. southwesternIdaho. Ecology 66:171-178. HOLTHUIJZEN,A. M. A. 1990. Prey delivery, caching, SMITH, D. G., ANDJ. R. MURPHY.1979. Breeding re- in PrairieFalcons. Con- and retrievalrates nesting sponsesof raptorsto jackrabbitdensity in the east- dor 92:475-484. ern Great Basin desert of Utah. RaptorRes. 13: KNICK,S. T, ANDJ. T. ROTENBERRY.1997. Landscape 1-14. characteristicsof disturbed habitatsin shrubsteppe SMITH,D. G., J. R. MURPHY,AND N. D. WOFFINDEN. southwesternIdaho (U.S.A.). LandscapeEcol. 12: 1981. between abun- 287-297. Relationships jackrabbit dance and ferruginoushawk reproduction. Condor KNICK, S. T., J. T ROTENBERRY,AND T. J. ZARRIELLO. 83:52-56. 1997. classification of Landsat the- Supervised SPSS, INC. 1997. SYSTAT 7.0 for Windows. SPSS matic in a semi-arid by mapperimagery rangeland Inc., discriminate Photogram- Chicago. nonparametric analysis. J. S. H. ANDERSON,AND R. OAKLEAF. metric Engineeringand Remote Sensing 63:69- SQUIRES, R., 86. 1993. Home rangesize and habitat-usepatterns of oil in KOCHERT,M. N., ANDM. PELLANT.1986. use nesting PrairieFalcons near developments Multiple J. Field Ornithol.64:1-10. in the Snake River Birds of Prey Area. Range- northeasternWyoming. lands 8:217-220. STEENHOF,K. 1987. Assessing raptorreproductive suc- In KORPIMAKI,E. 1984. Population dynamics of birds of cess and productivity,p. 157-170. B. A. Giron prey in relationto fluctuationsin small mammal Pendleton,B. A. Millsap, K. W. Cline, and D. M. populations in western Finland. Annal. Zool. Bird [eds.], Raptormanagement techniques man- Fenn. 21:287-293. ual. Natl. Wildl. Fed., Washington,DC. KOSTRZEWA,R., ANDA. KOSTRZEWA.1990. The rela- STEENHOF, K. 1998. PrairieFalcon (Falco mexicanus). tionshipof springand summerweather with den- In A. Poole and E Gill [eds.], The birds of North sity and breedingperformance of the buzzard(Bu- America, No. 346. The Academy of Natural Sci- PRAIRIEFALCON POPULATIONS 41

ences, Philadelphiaand the AmericanOrnitholo- U.S. DEPARTMENTOF THEINTERIOR. 1996. Effects of gists' Union, Washington,DC. military training and fire in the Snake River Birds STEENHOF,K., ANDM. N. KOCHERT.1982. An evalu- of Prey National Conservation Area. BLM/ ation of methods used to estimate raptornesting IDARNG Research Project Final Report. U.S. success. J. Wildl. Manage.46:885-893. Geol. Surv., Biol. Res. Div., Snake River Field STEENHOF,K., ANDM. N. KOCHERT.1988. Dietaryre- Sta., Boise, ID. sponses of three raptorspecies to changing prey VAN HORNE,B., G. S. OLSON,R. L. SCHOOLEY,J. G. densities in a naturalenvironment. J. Anim. Ecol. CORN, AND K. P. BURNHAM.1997. Effects of 57:37-48. drought and prolonged winter on Townsend's STEENHOF,K., M. N. KOCHERT,AND T. L. MCDONALD. ground squirrel demography in shrubsteppe habi- 1997. Interactiveeffects of prey and weatheron tats. Ecol. Monogr. 67:295-315. Golden Eagle reproduction.J. Anim. Ecol. 66: VAN HORNE,B., R. L. SCHOOLEY,AND P. B. SHARPE. 350-362. 1998. Influence of habitat, sex, age, and drought U.S. BUREAUOF LANDMANAGEMENT. 1988. Final en- on the diet of Townsend's ground squirrels, Sper- vironmentalassessment land Idaho mophilus townsendii. J. Mammal. 79:521-537. report, Army VAN AND B. National Orchard Area HORNE,B., P. SHARPE.1998. Effects of Guard, Training Multipur- armored vehicles on Townsend's pose Complex. Bruneau Resource Area, tracking by Range ground squirrels in the Orchard Training Area, Boise, ID. Idaho. Environ. 22:617-623. U.S. DEPARTMENTOF THE INTERIOR. Snake Manage. 1979. River VEKASY,M. S., J. M. MARZLUFF,M. N. KOCHERT,R. Birds of Prey Special ResearchReport to the Sec- N. LEHMAN,AND K. STEENHOF.1996. Influence of retary of the Interior.Bur. Land Manage., Boise, radio transmitters on Prairie Falcons. J. Field Or- ID. nithol. 67:680-690. U.S. DEPARTMENTOF THE INTERIOR. 1995. Snake River YENSEN, E., D. L. QUINNEY, K. JOHNSON,K. TIMMER- Birds of Prey National Conservation Area man- MAN, AND K. STEENHOF. 1992. Fire, vegetation agement plan. Bur. Land Manage., Bruneau Res. changes, and population fluctuations of Town- Area, Lower Snake River District Office, Boise, send's ground squirrels. Am. Midl. Nat. 128:299- ID. 312.