장음표시 사용
21쪽
spatiat autocorrelations by spatiat correlograms are useful to reveat habitat selection Datures of habitat speciali sis and generali StS. Line transect censu ses are osten used in avi an biology for monitoring popu
sus accuracy affect the usabili ty of these censu ses for the analysis of spatiat auto- correlation in the distribution os bird territories. Is spatiat autocorrelation can bedetected hom transect censu ses of territorial hircis. then this would offer an easy-to-use technique for Geld ornithologi sis. In the present paper I use transect data tore Veat spatiat autocorrelation. Spatiat autocorrelation on the data seis ari sed Domterritory mapping and line tran Sect censu Se S are also compared. In addition, changes of spatiat autocorrelation in the territory structure during the breed ing
Both spatiat correlograms and semi- variograms are osten used to displayspatiat autocorrelation. They show the fame information, hut tendencies arereVersed . In the present paper Ι prefer correlograms oVer Variogram s. hec ause the
22쪽
casual representation of the correlogranas eas ity demonstrates the negative and positive autocorrelations. Neverthel ess, under Special circumstances Semi-Vari O- granas may be more informative than correlogramS, Or ViCn VerSI.
Spatiat autocorrelation of 4 passerines is gi ven on Fig. l. Although the correlograms presented here cover a wide range of the spatiat Scales, the Values onthe lest fide of the figure are more important than values on the right at higher distance classes. The figure shows that the Sedge Warbier Acrocephalus SchoenobaenuS) formed sinali territories in patches. One km 0. 8-l. 8 km) longsections of reed bed showed an intensive clumping, but another 0. 2-l. 9 km longsections remat ned empty. Spatiat autocorrelation shows hi gh autocorrelation atthe first steps. Because of the clumping of Sedge Warblers' territories there is alii gli chance to find another territory in the close vicinity of any territory. Thecori elograna of the Reed Warbier Acroc haluS Scit Pacet S) Show no autocorre
23쪽
whole transect than the Sed ge Warbier. Although the habitat is not saturaled. thelii gli densi ty of Reed Warbier territories makes hi gh values of autocorrelation atlags 3-5. The autocorrelogram of the Reed Bunting Deriga Schoenicius) territories silow a speciat figure. It has not revealed autocorrelation at the first fewlags. It is explained by the greater territories of this species, whicli forna long nar-row territories along the Census rOute. i. e. in edge habitat of the marshland . Sincethe transect was handi ed as a continuotis set of 200 m long segments. one Reed Bunting territory osten occupied more than One of these SegmentS. Furthermore, Reed Bunting territories did not jo in to each other, So smali empty sections were mund at severat places. The smali territories of the Savi's Warbier Locustellia luscinioideo also fhowed sonae clum ping. but they distributed not as evenly asthe Sedge Warbier. The figures of the lalter two species were similar to eachother, hut the Savi s Warbier showed somewhat Smaller autocorrelations. The former resulis were based hi pon the territory maps. Single line transectcensu ses may also have the potentiat use in autocorrelation analysis. The mapping method resulis in territori es hereas the single line transect censuses resultin individual hirci observations. As hird territori es have so me considerable extention; it is difficult to say what part of the territory is necessary to u se in spatiat autocorrelation analysis. In the present stud y Ι characteri Zed each territory by apoini, which was chosen as the centre of the territory. Another possibili ty is tosearch for nesis, and choose nest positions inste ad of the centre of territori es. Although the lalter method is ad vantegotis in Some cases, it is time-consuming and many nesis may not be found even after substantiat labour. Single line transectcensu ses can handie single observations of territory-holding bird s. so the spatiat patieria reve aled may depend on the fact where in the territory was the given bird Observe l. This problem arises when territories are large, at hi glier densities. and at fine-scale Os Observations, e .g. when s mali distance classes are u sed in transect data. Figure 2 shows the dendrograna of cluster analysis based on the spatia lautocorrelations of J consecutive line transect censu ses of Sedge Warblers
Fig. 2. Agglomerative cluster analysis of autocorrelation structure sup to 6 lags) obtained by the territory mapping method lapping ) and 7 consecutive line transcct censu ses, denoted by the clate os census. Autocorrelation was calculated by the Moran 's I sor the Sedge warhler
24쪽
SPATIAL AUTOCORRELATIONS AMONG BIRD TERRITORIES
3). Autocorrelation values at hi gh lags may have limited relevance hecause of the decreas ing sample si Ze.) Most correlograms showed a minimum inflectionpoliat at lag 4. where no or litile negative autocorrelation was detected. Only the first two censu ses carried out in April showed a minimum value at lag 3 no auto- correlation in census ''April l 6 ), or non significant negative autocorrelation atlags 3 and 4 in census ' April 26 ). The lalter two censu ses were separated in thedendrograna of the agglomerative cluster analysis Fig. 2). Although a single line transect census may deieci more birds than the mapping method, especialty when non breed ing birds smigranis or unpared males)stay temporari ly in the area, they generat ly underestimate densities. The mappingmethod is a time-consum ing procedure, hut resulis in a good estimation of Vtrue densities. The line transect technique is rapi d. and it is easy to carry out . For this reason the mapping method is regarded as an 'absoluteV technique. but the lessi eliable line transect method is known as a 'relative ' method JARVINEN & V I SANEN l9Jβ). MOS AT and BALDI l 999) analyZed the fame data-set u sed in thepresent paper in order to evaluate the u semines of the Pinnisti line transectmethod in marshland habitat. For the Sedge Warbier the main beli data showed
25쪽
Fig. 3. Correlograms of the ahoundance of the Sedge Warbier obtained by the territory mappingmethod Mapping ') and by single line transecis denoted by the date of the census). Ordinate Morams Ι version Zd) spatiat autocorrelation coefficient. Significant autocorrelation coefficients PQ0. 05) are indicated by Open circles
26쪽
SPATIM AUTOCORRELATIONS AMONC BIRD TERRITORIES
Not only the number of territories changed over the breed ing SeaSOn se. g. WIENS i 969). hut the number of non- statio nary birds as well smigranis: ny ingacross the area, or Staying in the area for a few days: Roaterso ian maled males),which may modify the resulis of the line transect censu ses MOS AT & BALDIl99 β). Nevertheless, the breed ing population is difficult to estimate hec ause of predation. Interspecific territoriali ty may also modify the Single-Species palteria. For these reasons reseachers have to take into consideration these factors when specific questions are ad dres sed to spatiat ly related ecological problems. The mapping method is sensitive to measti re an average state of the breed ing season excluding the extreme Values of migration pe3ks irregular Roaters and low sing-ing activity, i. e. the low detectabili ty periods of the territory holder males. Forthis reason the mapping method seems to be the best technique is autocorrelationof the the spatiat population is examined, hut main treiads may also be revealed by single line transecis. Single line transect censu Ses seem to be u Sesul to demonstrate changes of spatiat autocorrelation during the breed ing Season, when autO- correlation is robust, and well-defined.
and his comments On the manu Script.
27쪽
Revi sed version received idth March. 2000, accepted fili June. 2000. publis hed Pth July. 2000
28쪽
Acta Mologica Academicie Scientiarum Hungariciae ψ6 I); ρρ. 2 33, 2000
HABITAT SELECTIONOU MIGRATIN G SED GE WARBLERS ACRO CEPHALUS SCHOENOBAENUS) AND MARSH WARBLERS A. PALUSTRIS) IN A
FelSOmialom u. 22, H-Z62J PecS. Hurigar Data from a ringing program in a Sohath Hungari an reed swamp. were used to analyse the ho-rigontal and vertical distribution os the juvenile, adult Sedge Warblers and Marsh Warblers during autumia migration. The hori Zontal spatiat patierra os the Juvenile and adult Sedge arblers showed cluniped distribution in the reed bed. One thousand three hundred and twenty- sive 35 R) juvenile and 203 2PR) adult Sedge Warblers were caught in three internal nets of the first line os misi-neis; which represenis ora ly l6 , of the nets. These 3 neis arestandisag in the 30 0 m wide Ione near open water. Juvenile Marsh Warblers' distribution vas clumped. but there was no significant irenis in the number os juveniles along the si ore open water gradient R2 α 0.0J). Aduli Marsh Warblers' distribution was rando in in the i cedhecl. The vertical spatiat distribution of the age classes were t significantly clumpcd sor both species. The investigation os 1 ood dispersion in a reed bed. the vegetation structure and the interspecific competition among reed warblers is require t to bellor understand the habitat se
l99J) and speiad the winter in Africa; solith of the Sahara DO SETT et a l. l 988). The habitat selection is one of the most important mechanis ins in the ecological segregation os bird populations. The spatiat distributions os passerine species during the breed ing and the wintering BARLEIN l 983) seasons have been studi ed in more delati than during migration BACCETTI l98 PAMBOUR l 990. HONTA & LlTERA K l99 ). Food avallabili ty and vegetation structure are known
29쪽
In one of our former studies - with less individual number and different statisticat methods, howeVer - we furveyed the area distribution of the Sedgewarbier in relation to the age sGYURACZ R BANK l99J). In this study. data Doma long-term ringing program in a solith Hungari an reed Swamp, were u Sed to compare the hori Zontal and vertical distribution os juvenile and adult Sed ge Warblers and Marsh Warblers during posthreeding migration. We checked two hypotheses l) The area distribution of the two warbier species cloes not differ in themigration period. 2) The area distribution of the juvenile and adult hirds cloes not differ in the migration period os ei ther species.