So far reported unstable correlations between environmental settings and doline geometry require a revision of doline delineating methods and morphometric analyses that follows from such input data. We examine the reliability of the most accessible data sources for doline delineation, namely basic topographic maps at a scale of 1:5,000 and ortho-rectified aerial photographs (digital orthophotos) of 0.5 m resolution. Their reliability has been quantified in the Dinaric karst (Krk Island, NE Adriatic Sea). The adequacy of most commonly used morphometric parameters (viz. elongation and circularity) and their variations applied in previous studies are mathematically verified and practically tested. The simulation of an individual method behaviour performed on polygons with increasing circularity and elongation demonstrated their ranges of validity. The efficiency of a particular method was verified by the application to the real dataset of dolines. The optimal variation of an individual parameter (circularity = Pcc/Pd; elongation = of the best fitting ellipse) is identified and suggested for use in order to provide uniform results and increase the comparability of future morphometric analyses. Asymmetry (the distance between the centroid of doline ground plan and the centroid of the largest inscribed circle) as an additional parameter for describing doline ground plans has been developed and is proposed for further use.