The distribution and abundance of the rook corvus frugilegus L. as influenced by habitat suitability and competitive interactions.

Abstract

Rooks (Corvus frugilegus) are colonially breeding corvids found in most agricultural landscapes. Colonies in the County Durham area tend to be clustered at distances up to 500 m, but otherwise show little pattern in terms of spacing or size. Colony size was comparable between sites as changes in colony nest counts were allowed to stabilise before the whole area was surveyed. When measuring nest build-up at a sample of colonies in 1996, no further significant increases occurred after 9th April. The spatial size distribution of colonies was maintained between years. The distribution and size of breeding colonies is modelled in relation to the interaction between the spatial distribution of the foraging habitat and potential intraspecific competitors, with the identification of the distance over which this interaction is strongest. The satellite derived habitat data used for the modelling were part of the ITE Land Cover Map of Great Britain. However, their correspondence with ground reference data was found to be severely lacking. Thus, for modelling the availability of nesting habitat, OS woodland data were used as these identified more of the extant rookery sites, whilst the ITE data were retained for quantifying the foraging habitat. Logistic regression showed that the distribution of colony sites was influenced by the availability of woodland blocks large enough to hold a colony, proximity to roads and buildings, and by the amount of pasture within 1 km. Other suitable sites with these characteristics remained unoccupied within the distribution. Partial Correlations showed that interactions between the spatial distribution of the foraging habitat and competitors influenced colony size at distances up to 6 km, suggesting their effect outside of the breeding season. The multiple regression model built with variable values for this distance explained 31% of the variance in colony size. When applied to the potential breeding sites identified using the logistic regression, most sites still remained suitable. This suggests the distribution is not saturated and that limited availability of breeding habitat is not the cause of the nesting aggregations. The broad correlation of Rook abundance to foraging habitat and potential competitors corresponds to an ideal free distribution of individuals across colony sites. This is supported by models of Rook numbers in relation to parish agricultural statistics produced by MAFF. These again show the importance of pasture as a probable foraging resource, and how pasture quality could be important to Rook numbers. The models also supported the ideal free predictions of spatial variation in Rook abundance in relation to habitat, and the response of colony sizes to temporal change in habitat quality.