J. Raptor Res. 27(2):102-105 ¸ 1993 The Raptor ResearchFoundation, Inc.

NEST-SITE HABITAT SELECTED BY COMMON BUZZARD (Buteo buteo) IN SOUTHWESTERN FRANCE

CHRISTINE HUBERT Centrede Rechercheen Biologiedu Comportement,U.A. C.N.R.S. 664, UniversitdPaul Sabatier, l l8, route de Narbonne, 3•062 Toulouse,France

ABSTRACT.--I studied the habitat characteristicsassociated with nest sites of Common Buzzards (Buteo buteo)in southwesternFrance during threeyears. Nineteen nest sites were comparedto 60 randompoints in buzzardnesting habitat using21 quantitativehabitat variables. Most nestswere in large pinesor oaks and in the upper 2/3of the tree at a mean heightof 13.4 m. It appearedthat buzzardsselected nest sites in mature woodedareas with easyaccess, rather high in the tree, and closeto woodlandedges.

Selecci6nde sitiosde nidificaci6npor Buteobuteo en el suroestede Francia RESUMEN.--Estudi• las caracteristicas del hfibitat asociado a los sitlos de nidificaci6n de Buteo buteo en el suroestede Francia durantetres aftos.Se compararon19 sitiosde nidificaci6ncon 60 puntosescogidos al azar en el hfibitatde estaespecie; se usaron21 variablescuantificables del hfibitat.La mayoriade los nidosse encontraronen grandespinos o roblesy sobreel terciosuperior del firbol,a una altura media de 13.4 m. A1 pareceresta especie selecciona sitios de nidificaci6nen fireasde firbolesmaduros, de fficil accesoa la cima del firboly cercanosa fireasabiertas. [Traducci6n de Ivan Lazo]

Although many nest-sitepreferences have been were made to eachnest to seeif buzzardswere attending describedfor the Common Buzzard, few studieshave or defendingnest sites, and later in the springin order to seeif there were broodingbuzzards or nestlings. attemptedto determinethe decisivefactors involved All nest sites were plotted on a 1:25000 topographic in its nest-site selection. These studies have con- map. A nest site was definedas the nesttree plus a 0.04- cludedthat tree nestssimply needto be adequatein ha circular plot (11.3 m radius) centeredon the nest tree size and shape. No previousstudies attempted to (Titus and Mosher 1981). Within the nestsite, all woody describenest-site habitat of the buzzard quantita- plants over 1.65 m were tallied accordingto species,di- ameter at breastheight (dbh) and whetherthe predomi- tively. nant foliagewas part of the overstoryor understory.Sep- arate size categorieswere createdfor understoryplants MATERIALS AND METHODS and overstorytrees. In sucha way, 21 quantitativehabitat Study Area. I conductedmy studyin the forestof Bou- variables(Table 1) were either measureddirectly or cre- conne (2300 ha) located near Toulouse (southwestern ated by aggregation.Height measurementswere made France) which is surroundedby cultivationand villages, usinga clinometer.Distances were measureddirectly on and lies between 200 and 300 m elevation. Dominant the field when short (<50 m), or measuredfrom the map. overstorytrees are oaks, including Quercuspedonculata, Percentageof vegetationcover was simply visually esti- but also Q. sessilifloraand Q. lanaginosa,and introduced mated. pines(Pinus sylvestris and P. pinaster).The understoryis If the plot fell at the edge of a field or other forest mixed with hornbeam(Carpinus betulus), chestnut (Cas- opening,no correctionswere made, this opening being tanea satira), ash (Fraxinus excelsior),maple (Acer cam- consideredas part of the plot. Altitude was not considered pestre),lime (Tilia platyphyllos)and wild servicetree (Sor- here because of the flatness of the forest, nor was the busaria). Undergrowthis composedof hawthorn(Crataegus exposure. spp.), blackthorn(Prunus spinoza), broom (Sarothamnus Available NestingHabitat. Randomsampling was used scoparius),heath (Erica spp.), holly (Ilex aquilifolium)and to estimateavailable nestinghabitat. A numberedgrid bramble(Rubusfruticosus). The forestis subjectto logging: (grid squaresof 250 x 250 m) of 226 pointswas overlaid oaksare cut when 160 yearsold and pineswhen 80 years on the forest.Sixty pointswere drawn at random (Titus old. Becauseof logging,the forestis dividedinto numbered and Mosher 1981). Each randompoint was plottedon the sectionsdelimited by large pathways. map. Once the approximatesite was locatedin the woods, Nesting Habitat Description. Buzzardnests were lo- the nearesttree was made the centerof the randomplot cated in the winters of 1989, 1990, and 1991 through A tree was used to remain consistent with a nest-site sam- systematicfoot searchesof % of the forest.Frequent visits ple plot. Except for the nest-tree specificvariables, the

102 JUNE 1993 COMMONBUZZARD NESTING HABITAT 103

Table 1. Samplemeans and standarddeviation of nestsite and randomplots and significantdifferences between randomand nest-siteplots (ANOVA (F) and Mann-Whitney (U, z) valuesare given).

NEST SITE (N = 19) RANDOM(N = 57) VARIABLES (RANGES) (RANGES) TEST P VALUES Canopy height (m) 20.6 + 0.83 16.98 + 0.6 **a P-- 0.002 (12.5-23.4) (10-25.4) F(1,75) = 9.95 Distance to nearestopen water (m) 901 + 162 805 + 46.2 P = 0.43 (20-2425) (20-1750) F(1,75) = 0.61 Distanceto nearestwoodland edge (m) 211 + 46.1 480 + 40.8** P = 0.0007 (14-887) (0-1000) F(1,75) = 12.6 Distanceto pathways (m) 67.4 + 7.5 54.2 + 5.8 P = 0.13 (16-137) (0-162) U = 418, z = -1.5 % of canopycover 35.3 + 3.4 37.37 + 2.1 P--0.617 (15-60) (10-75) F(1,75) = 0.25 % of understorycover 34.5 + 3.6 32.8 + 1.7 P = 0.94 (20-60) (20-60) U = 536, z = -0.06 % of ground cover 54.7 + 6.2 40.79 + 3.3 P = 0.059 (10-95) (10-100) U = 385, z = -1.88 Number of speciesof overstory 1.42 + 0.12 1.56 + 0.07 P = 0.27 (1-2) (1-2) U = 461, z = -0.96 Number of speciesof understorycover 2.68 + 0.32 3.05 + 0.14 P = 0.22 (1-5) (2-6) F(1,75) = 1.54 Number of ground species 4.58 _ 0.40 4.61 + 0.24 P = 0.94 (2-8) (2-8) F(1,75) = 0.005 Number of overstorytrees 17.2 + 2.01 27.47 + 1.90'* P = 0,004 (1-35) (1-68) U = 308, z = -2.80 Number of overstorytrees <25 cm dbh 10.7 + 2.36 23.63 + 2.21' P = 0.002 (1-35) (1-68) U = 288, z = -3.03 Number of overstorytrees 26-50 cm dbh 5.42 _+ 1.02 3.54 + 0.52 P = 0.035 (1-18) (1-11) U-- 367, z = -2.09 Number of overstorytrees >50 cm dbh 1.05 + 0.24 0.3 _+ 0.07*** P = 0.0003 (1-3) (1-3) U = 288, z = -3.6 Number of understorystems 1-4 cm dbh 18.21 + 6.33 40 + 13.9' P = 0.01 (0-124) (1-800) U-- 338, z = -2.4 Number of understorytrees 5-8 cm dbh 16.05 + 2.32 14.65 + 1.75 P= 0.36 (4-39) (1-72) U = 465, z = -0.91 Number of understorytrees >9 cm dbh 15.74 + 1.95 11.25 + 7.71 P = 0.04 (5-32) (1-38) U = 373, z = -2.0 DBH of the nest tree 52.85 + 2.76 (33-73) Hetght of the nest tree (m) 20.6 _ O.83 (12-23.4) Height of the nest (m) 13.39 + 1.02 (4.1 _ 23.2) a * = p < 0.01; ** = P < 0.001; *** = P < 0.005.

samplingvariables were the same as that of thenest site. generally allowed the sampling of the total forest area The criterion for acceptinga random plot was that the without preselecting"representative" or "typical" areas plot must be within a forestedarea with a canopyheight (Green 1979 in Titus and Mosher 1981). greaterthan or equalto 10 m. This excludedsome habitats Statistical Methods. Parametric and nonparametric in which buzzardsdo not nest (young plantation areas) statisticswere usedwhere appropriate.Univariate ANO- but included some areas where nesting was unlikely VA was conductedon the two groups (buzzard nest site (copsewoods,partly cut areasof timberwoods). This scheme and random samples)for 6 of the 17 variableswhich had 104 CHRISTINE HUBERT VOL. 27, NO. 2

Table 2. Proportion of oaks and pines in the forest of pines(over 26 cm dbh) as determinedfrom the ran- Bouconne(from 57 randomplots). dom sampling versus the proportion of nest trees used (N = 57, Z -- 1.318, P = 0.19). Numerous % studieson buzzards show a preferencefor a tree % OVERSTORY OVERSTORY % TOTAL species.Oak is generally selectedin France (Roche TREE TREES <25 TREES >26 OVERSTORY 1977, Nore 1979). In Finland, 66% of buzzard nests SPECIES cm dbh cm dbh TREES are in sprucePicea sp. (Solonen1982). In Germany, Pine 2 56 9 70% of the nestsare in beechFagus sp. (Rockenbauch Oak 98 44 91 1975 in Bayle and de Ruffray 1980). In the United Kingdom, pines are sometimesused (Dare 1961, Picozzi and Weir 1974) but rock sites are also se- a distribution close to the Gauss curve. For the 10 other lected(Brown 1976, Dare 1989). In this study,the variables (showing nonnormal distributions), Mann- selectionof nest-tree speciessimply reflected the Whitney testswere conducted. availability of large specimensof availablespecies. Buzzards also nestedsignificantly nearer to edges RESULTS AND DISCUSSION of woodlandsthan expected(Table 1). Althoughone I found a total of 33 nests of which 23 nests were nestwas 800 m away from an edge,50% of the nests occupiedat leastonce by a buzzard pair during the were lessthan 200 m from an edge.The importance three years.One nestwas occupiedtwice by a pair of woodland edgeshas already been mentionedfor of EuropeanSparrowhawk (Accipiter nisus), one nest buzzards (Joenson 1968, Nore 1979, Kostrzewa by a pair of BootedEagle (Hieraaetuspennatus), and 1987), and is explained by the fact that buzzards one nest by a pair of Goshawks (Accipitergentilis). hunt mainly at the woodland edge or on the sur- The other nestsshowed no evidenceof occupation roundingland (Joenson1968). Thiollay (1972) em- or defenceby any species.Four actively defended phasizedthat buzzard densityis correlatedwith the buzzardnest trees were cut downby foresters.Hence lengthof contactbetween woods and openland: 600- only 19 buzzardnest sites were characterizedquan- 900 m of edge per pair. The associationbetween titatively. buzzardsand forestopenings could be evenstronger No nestswere foundin areaswith canopyheights when buzzardsnest on hedgerows(Roche 1977) or below 14 m (Table 1). The mean height of nests on pylons (Melde 1983). Moreover, closenessto was 13.4 m. Most nests(52%) were in the upper 2/3 woodlandedges facilitates nest accessibility(Roche of the tree, with 80% in the upper half. Buzzards 1977). nestedin mature treesmore than expected,based on Buzzards did not select nest sites in relation to the fact that most nests(89%) were in trees over 40 openwater; the distanceto pathwaysdid not influ- cm dbh (Table 1). The marked preferenceto nest ence the buzzard in its nest-site selection. Neither in mature plotscould be due to the fact that mature percentagecover (whatever the story) nor the num- treesallowed high nest placement, shelter from pred- ber of species(whatever the story) influencedselec- ators, and also gave a safebase to the nest (Solohen tion of the nest-site (Table 1). 1982). Most of the nestswere at the sameheight as Based on these results, I concludedthat Common understorytrees which gave additional protection Buzzard selectedmature woodedareas with an easy after the appearanceof the foliage (Morris et al. access,rather high in the tree itself, and closeto the 1982). Moreover, a low density of overstorytrees woodland edge. It could be interestingto compare around the nest site facilitatedaccessibility and vig- nest-site habitat of the Common Buzzard with the ilance. ecologicallysimilar Red-tailed Hawk (Buteojamaz- All speciesof overstorytrees were used as nest censis; Cramp and Simmons 1977): nest-site re- trees,but nestsoccurred slightly more often in pines quirementsof both speciesappear to be rather sim- (11) than in oaks (8). This trend dependedon the ilar (Orians and Kuhlman 1956, Titus and Mosher proportion of pinesand oaks countedin the random 1981, Bechard et al. 1990). plots (Table 2). Whatever the diameter,pines rep- ACKNOWLEDGMENTS resentedonly 10% of the total overstorytrees of the P. Winterton kindly improvedthe English.I acknowl- random plots. A chi-squaretest showedno differ- edge with gratitude V. Bretagnolle,A. Gallo, J. Lauga encesbetween the proportionof large oaksand large and M.J. Bechard for their commentson earlier drafts. JUNE 1993 COMMON BUZZARDNESTING HABITAT 105

LITERATURE CITED Neue Brehm Bficherei, Wittenburg Lutherstadt, Ger- BAYI•E,P. AND P. DE RUFFRAY. 1980. Analyse biblio- many. graphique de la Buse variable Buteobuteo. Bull. Nat. MORRIS,M.M.J., B.L. PENAIq,R.E. LEMMONAND D.M BIRD. 1982. Characteristics of Red-shouldered Hawk O.N.C. 40:11-15. BEeHARD,M.J., R.L. KNIGHT, D.G. SMITH AND R.E. (Buteolineatus) nest sites in southwesternQuebec. Can Field-Nat. 96:139-142. FITZNER. 1990. Nest-sitesand habitatsof sympatric hawks (Buteospp.) in Washington.J. Field Ornithol. NORE,T. 1979. Rapacesdiurnes communsen Limousin 61:159-170. pendantla p•riode de nidification,Buse, Bondr•e, Mi- lan noir, Busard St. Martin et cendr& Alauda 47:183- BROWN,L. 1976. British birdsof prey. Collins,London, 194. U.K. ORIANS, G. AND F. KUHLMAN. 1956. Red-tailed Hawk CRAMP, S. AND K.E.L. SIMMONS. 1977. The birds of the western palearctic.Univ. Press,Oxford, U.K. and Great Horned Owl populationsin Wisconsin.Con- dor 58:371-385. DARE,P.J. 1961. Ecologicalobservations on a breeding populationof the CommonBuzzard Buteobuteo with PICOZZI,N. AND O. WEIR. 1974. Breedingbiology of particularreference to the diet and feedinghabits. Ph.D. the Buzzard in Speyside.Br. Birds 67:199-210. Dissertation, Univ. of Exeter, U.K. ROCHE,J. 1977. Un recencementde Busesen plaine de 1989. Aspectsof the breedingbiology of the Saone; quelques donnilesconcernant la nidification. Buzzard Buteo buteo in north Wales. Naturalist 114: Jean-le-Blanc 16:48-63. 23-32. SOLONEN,T. 1982. Nest-sites of the Common Buzzard (Buteo buteo)in Finland. Ornis Fenn. 59:191-192. JOENSON,A.H. 1968. An investigationon the breeding THIOLLAY,J.M. 1972. Ecologie d'une population de populationof the Buzzard (Buteobuteo) on the island Als in 1962 and 1963. Dan. Ornithol. Foren. Tidsskr. rapacesdiurnes en Lorraine. La Terre et la Vie 2:116- 176. 62:17-31. TITUS, K. AND J.A. MOSHER. 1981. Nest-site habitat KOSTRZEWA,A. 1987. Quantitative analysesof nest- habitat separationin the CommonBuzzard (Buteobu- selectedby woodland Hawks in the central Appala- chians. Auk 98:270-281. teo), Goshawk (Accipitergentills) and Honey Buzzard (Pernisapivorus). J. Ornithol.128:209-229. MEnDE, M. 1983. Der Mafisebussard(Buteo buteo). Die Received8 December1992; accepted1 March 1993