Besides the scientific interest in the quantification of bird migration, there is an increasing need to quantify bird movements for the assessment of bird collision risk with artificial structures. In many environmental impact studies, the radar method is used in an inappropriate manner. The processing of echoes consists often of counting blips within defined screen fields, and the surveyed volume is estimated without reference to the detection probabilities of different ‘target sizes’ (radar cross-sections). The aim of this paper is to present a procedure to quantify bird migration reliably using radar by stating the theoretical requirements of every single step of this procedure and presenting methodological solutions using our own radar data from extensive field studies. Our methodological solutions can be applied to various radar systems, including widely used ship radar. The procedure presented involves discriminating the echoes of birds and insects and estimating the different detection probabilities of differently ‘sized’ birds (radar cross-sections). By ignoring the different detection probabilities, density estimations may be wrong by as much as 400%. We fear that quantification of bird migration and predicted bird numbers affected by collisions with artificial structures are in many cases based on unreliable estimates.