Realm: Terrestrial
Climate: Temperate
Biome: Temperate broadleaf and mixed forests
Central latitude: 55.716670
Central longitude: 13.333330
Duration: 57 years, from 1953 to 2009

1210 records

39 distinct species

Across the time series Fringilla coelebs is the most frequently occurring species

Methods

The same territory mapping method as used by Enemar et al. (1994) was applied also during the additional seventeen years. For a recent general description and discussion of the method we refer to Bibby et al. (2000). The method is also called spot mapping? after William (1936) and it is similar to the method used in the Common Bird Census in the UK (Marchant et al. 1990) and the Breeding Bird Census in North America (Robbins 1970). The observer walks slowly through the plot putting down all bird observations in their proper positions on a map paying particular attention to recording simultaneous presence of males in adjacent territories. Species maps are then constructed from the visit map records. The species maps are used to evaluate the number of distinct clusters of records which are believed to represent territories. The method is of course prone to errors both when recording the birds in the field and when evaluating the number of territories from the species maps. Furthermore the number of territories is never stable throughout the breeding season: some birds die some fail to breed successfully and leave their territories and some arrive late and establish new territories. However for our analysis it is not essential to know the exact number of territories in every particular part of every breeding season. Instead standardization (recording birds and evaluating the species maps in the same way every year) is the key to reliable comparisons and we did our best to achieve this. We made ten visits per year in all but two years (nine visits in 2007 and 2009). They were distributed from April through June (always after 9 April and before 18 June) but with variation according to the arrival of spring. All surveys were made by the same person one of the authors (Ann Mari Thorner). Since AMT also carried out all surveys in 1987?1992 there is now a series of no less than twenty-one years with the same observer. This ensures that the new results are comparable with the data from before 1993. Each visit took between two and three hours so the minimum annual survey effort was twenty hours. A cluster of records was accepted as a territory if it contained at least three records (for a few very late arriving species two records was considered sufficient; cf. Svensson 1978). Adjacent clusters were accepted as different territories either if they were distant enough from each other (in relation to typical territory size) to make it likely that they belonged to different males or if they contained simultaneous records (a male recorded in both at the same time). A special study on inter-specific competition with a large number of nest-boxes was conducted in the valley in 1963?1966 (Enemar et al. 1972). The populations of Great Tit Blue Tit and Pied Flycatcher were artificially enhanced these years. We removed the effect of the experiment by reducing the numbers to a proportion equal to the mean proportion in the four years before and the four years after the experiment. The number of Great Tits was reduced from 9 10 19 16 pairs to 6 6 12 10 respectively. The number of Blue Tits was reduced from 6 4 7 5 pairs to 3 2 4 3 pairs. The number of Pied Flycatchers was reduced from 15 18 18 13 pairs to 4 5 5 4 pairs. These modified values are used in all calculations in this paper. Species turnover rate between two different years was calculated by dividing the total of new and lost species with the number of species in year one plus the number of species in year two. This gives the value one hundred percent (that is complete) turnover when no species are common between the two compared years. This is different from the turnover calculation by Enemar et al. (1994) where the sum of new and lost species was divided by the mean number of species in the two years. The latter method gives one hundred percent turnover when half of the species have been lost and replaced with new ones (the methods are identical in principle but give results that differ by a factor of two). In the same way we calculated turnover between two years different number of years apart (from adjacent years to a maximum of fifty-six years apart that is 1953 and 2009). Altogether forty-one small passerine species were recorded as territorial in at least one of the fifty-seven years. Eighteen of these species were keeping territories in all or almost all years (the first eighteen species in Appendix 1 also listed in 33Table 1). These eighteen species are called regular? in this paper. The final twenty-three species of Appendix 1 that is those with zero values in many or almost all years are called non-regular species. Eight of these latter species were recorded in less than four of the fifty-seven years and these species are not at all tested for population change. For the remaining fifteen non-regular and the eighteen regular species we calculated and compared the average number of territories during the first forty and the last seventeen years respectively. The difference between the two periods for the regular species was tested using a two-tailed t-test (Microsoft Excel) and the logarithms of the original values (with a small number of zeros replaced with the value 0.1 to permit calculation of logarithms). The differences for the non-regular species were tested using frequencies namely the number of years with different number of territories (i.e. with 0 1 2 3 and so forth territories). When the number of frequency classes was larger than four frequencies were pooled to form exactly four classes the same ones for both periods giving a 2?4 cell matrix. The Fisher exact test was then applied using the Vassar Stats package (home page of Vassar College Poughkeepsee NY). For the regular species we also calculated the trends across all fifty-seven years using the logarithm of abundance and standard linear regression (Microsoft Excel) also here with zeros replaced with the value 0.1. The same method was used when we calculated trends for parts of the full time series in these species. When calculating correlations between species we used the Spearman rank correlation module of STATISTICA (version 6.1; StatSoft?). All cover values are measured from the vegetation measurement frame which is 1 meter by 1 meter and partitioned into a grid of 100 10 cm by 10 cm squares. Cover is measured by counting the number of 10cm squares that are occupied by the foliage canopy of a particular plant species or by the soil disturbance leaf litter etc. Portions down to the 0.1 of a 10cm square are also measured.? Unit of abundance = IndCountInt, Unit of biomass = NA

Citation(s)

Svensson, S., Thorner, A. M. & Nyholm, N. E. I. (2010) Species trends, turnover and composition of a woodland bird community in southern Sweden during a period of 57 years. Ornis Svecica 20, 31–44.