BULT H (1995) Trefclus: a model to check species maps. LIMOSA 68 (1): 29-34.

In large-scale breeding bird censuses problems arise when species maps are analyzed according to the guidelines of the SOVON breeding bird monitoring project (BMP, van Dijk 1993). Due to the smaller number of visits to the plot (four to nine), oversized territories arise (Fig. 1) when the second guideline, i.e. a maximum number of registrations in a minimum number of territories, is strictly applied. One solution is to pose a maximum to the size of the territory (e.g. 200 m cross-section for most woodland passerines). In order to check the validity of such an analysis a computer model ('trefclus') based on the registration efficiency of each species ('trefkans, tk', Kwak & Meijer 1985) was developed. The registration efficiency is the probability to detect territorial behaviour when visiting a territory under standardized fieldwork conditions (early morning census, fair weather, 3-4 minjha). The model was applied to 48 maps of common (at least nine territories/plot) species with unequivocal territorial behaviour in three study plots. The majority of the maps (87%) was in accordance with the model, provided that the maximum territory size was restricted (ef Great Tit, Fig. 2). The density of two maps (4%) appeared too high and required re-analysis, whereas the density of four maps (8%) of two species (Willow and Crested Tit) appeared too low. These deviations could be corrected by taking into account that the registration efficiency had been above average during the main survey period (Fig. 3). Furthermore, in both plots the Crested Tit featured among the most abundant passerines and its high density presumably favoured a further elevation of the registration efficiency (Fig. 4). It is concluded that the model offers an unbiased possibility to compare and check the validity of species maps based on a limited number of visits.

[free pdf] [dutch summary]