On the premigratory dispersal of the Lesser Kestrel Falco Naumani in Spain
Authors: Pedro P. OLEA
Published: Volume 48(2), December 2001. Pages 237-241.
Original Title: Sobre la dispersión premigratoria del Cernícalo Primilla Falco Naumani en España
In this Forum I analyse a paper by García Fernández (2000) that contains a number of methodological problems and biases which may have affected its results and conclusions. First, it is concluded that the estimate of the size of the Lesser Kestrel population provided by García Fernández (2000) is not reliable because: a) he extrapolated census data for two subareas summing up 124 km2 only to a total area of 575 km2 and b) the spatial distribution of the Lesser Kestrel (LK hereafter) in this area was very aggregated around roosts (3 km radius, 28 km2), that held on average 19 times more birds than areas further from roosts (Fig. 1) and were heterogeneously distributed in the study area. In addition, the confidence interval given by García Fernández (2000) for his estimate (747-1.587 birds) appears to suffer from calculation mistakes: for a mean bird density of 2.03 birds/km2 xi mum number of birds). Contrary to García-Fernández's suggestions, these latter data are not consistent with the phenology of the species (postfledging dispersal and migration), nor are they supported by a study which monitored a large roost located in the same study area (Table 1). Although further studies are required, it is suggested that the results of García Fernádez (2000) could have been an artefact of the method he used to estimate bird numbers, as the quantification of the density of birds in two small subareas (some tens of km2) could have been seriously affected by presumable local movements of the birds within the study area. Third, the author carried out an analysis of ringing recoveries to examine the premigratory dispersal of the LK in Spain using recoveries in the months of July, August and September. However, 18% of the recoveries analysed occurred in September, when migration takes place (peaking in mid-September). Therefore, the author has analysed mixed movements due to both premigratory dispersal and migration (September), a fact that could have biased the median dispersal distances and the median altitude of the ringing recoveries. I give concrete suggestions aimed at improving the census methods as well as to clarify the results obtained, and conclude that further studies are required to ascertain what are the causes of the large number of LK seen at the end of the summer at the northern border of their Spanish distribution area.