장음표시 사용
111쪽
abundance of breed ing hirds could be expla ined by very law variabies characteri sing farin land . Among these. the most important were Vari abies describing non-farmed. patchily distribu ted habitatS - woods, shmbs or marshes stheir percent- age cover or relative tengili). In Poland research based on precisely meas ured features of land scape structure and entire communities of hreed ing birds isneeded. Admitted ly, sonae studies have considered differen ces in the richness of communities in relation to land scape structure, hut the characteri sation has ei therbeen a pure ly de Scripti Ve One, Or presenting Only the percentage sitares of differ
The aim of the study was to reveat the relations hips belween farnatandstructure and the abundance of breed ing birds, i. e. how do the amount of treeco Ver, meadows, the densi ty of habitat edges and the diversity of agri culturalland affect the structure of avi an communitieS.
112쪽
BIRDS AND AGRICULTURE LANDSCAPE IN POLAND
0002038 22029 Cropcrop sia ha)4440608950040245
to the west there are large complexes of pine forest. In the Gields there are stili mostly deciduoustrees, localed mainly in wet paris of the area and along the roads. The comparative characteristics of the study plois are gi ven in Table l.
weather and hird activi ty. For more exact mapping. potnis os reference were designated and markedin woods with a s hortage of natural orientatiori potnis. The numbers of pairs of Buteo huteo. Strisia luco, corvids and SturnuS Vulgaris were determine i ora the basis of the numbers of nesis sinund. CucuiuS cianorus was excluded frona the analyses due to iis non- territorial breed ing bellavi our whicli mnes the densi ty estimation uia reli able. The biriis bree ling on the Study plots were groia pedinto nesting-guilds according to TRYIANO SKI l 999). Statistical analysis os relations hips hetween land scape structure and abundance of birds onthe study plots was carried out using simple regression and sorward Stepwi se multiple regression. The dependent variabies were the number of species and the densi ty of bird populations saccordirigio hole communi ty or to nesting-gui id s), while the independent Vari abies were the ones describ-ing plot structure. The Vari abies expressed as proportions were transformed with the arcsine sx)function for normali sed values of the proportional type ZAR l 984). Aster transformations the distribution os ali variabies did not differ frona normal Olmogorov-Smii nox test, P 0. 05). Ali statis
113쪽
Tahle 2. Structure parameters os breeding bird communities os studi ed plois
114쪽
BIRDS AND AGRICUL RE LANDSCAPE IN POLAND
l able 3. Ρearson correlation coes cients bet een 'bird ' on 'habitat' parameters. Explanations os
Densi ty of lowlynesting Species
115쪽
most important for the number of species in this gui ld. but together with grass-land explaining 90R of the variabili ty. All the resulis are collected in Table 4. Summing up the resultS of the Step-wi se regression are very similar to the resulisos correlation coefficient analysis and they indicate a decisive role of wood habitat s for explaining the between-plois variation of hird densit y and abundance. The analyses presented Shows that in most cases sexcept the densi ty of ground -nesters) the relations hips belween ' habitat' variabies and species number
Tahle 4. The resulis of the forward ste pwi se multiple regression of 'bird ' on 'habitat' variabies Vari ableCoefficient
116쪽
BIRDS AND AGRICULTURE LANDSCAPE IN POLAND
117쪽
able belween-plois variation in bird densi ty, as weli as number of Species. Our resulis stlowed that differen ces in the hird communities. as weli as in the nesting-guilds in the agri cultural land scape. may be expla ined by reference to a s mallia umber si 2) of Vari abies characteri sing the land scape structure. It is clear Domthe resulis that tree cover plays a decisive role in the forming of bird communities in agricultural land scapes. BONGIOR NO l 982), SCHIFFERLI et al.
& PAR l 994) obta ined similar resultS. However the resulis frona Ireland LYSAGHT l 989) do not poliat to a positive relationship between the overati densi ty of linear hedges and the number of SpecieS, and population densi ty. Calculated on the basis of data mund in the cited work. the Spearman 's rank correlation coefficients, i s were e Vela negati Ve - withrespective values of -0.lf and -0.20 P 0. 05). This unexpected result may bellia ked to limited variation in the densi ty of hedges from l 22-l P in ha). The fame explanation may be gi Vela for the sui prising result obtained by BULL et til. l9J6). In this stud y the remo val of l 3 of the hedges in an area had no noticea bienegative effect on the overati population densi ty, or species richness of the bird communities despite the obviolis decline in species density e. g. in hole-ne Stet S. AS a consequence, analyses of Variation in bird abundance could be interpreted at two spatiat scales: sa) scale of a single element of land scape, and b)
118쪽
BIRDS AND AGRICULTURE LANDSCAPE IN POLAND
time for forest species to adapt to altered habitats O 'CONNOR & SI RUBB l 986).Very long transition frona the forest to Geld system may also explain why up to80R of farnaland avitatina in Great Britalia represent forest species, while it istes s than 30 50R in Poland . In conclusion the most important land scape element affecting bird richnessand abundance in fari land is the tree cover percentage. This variabie explain s63R of the variabili ty in bird species numbers and 85R of the total densi ty.
Among ne st-guilds Only the density of gro und-nesters was not correlated to any of the measti red 'habitat' variabies. The other nesting-guilds may be predicted in relation to tree cover and the determination coefficient regat ding to total densi tyamouia ted to O. 94-0. 99, and regarding to species number 0. 48-O. 90. The resulis stlowed essentiat importance of marginat woodland habitais for conserving hii drichness and abundance in farioland of West Poland.
Ac nowledgemenIS - we are grate sul to the sarmers who granted us regular access to theirland and supplied valvabie information on sarmi and management. We are grate sul to Z. KOSINSKI, A. SURMACKI. and especialty R. I. FULLER and three anonymous reviewers for their hel psul comments on an earlier drast. The study was partly mund by the Committee for Scientific Research