Among different dead wood types, large snags have a particular ecological relevance for biodiversity in forest ecosystems. For both research and management purposes, rapid and cost-effective inventories of large snags are required. Due to the great variability within stands and across the landscape, field recording of large snags is labour-intensive and expensive, if adequate sample sizes have to be ensured. We present a new method enabling efficient mapping and quantification of large snags by coupling colour infrared aerial photographs and a geographic information system (GIS). The method is validated by comparing the results with the data assessed by field methods in four spruce-dominated mountain forests in Switzerland. The different steps for implementing the method are: (1) stereoscopic interpretation of aerial photos for snag detection; (2) scanning and production of orthophotos; (3) geo-referencing and integration of the orthophotos and other data layers into a GIS; (4) digitisation of detected snags and drawing up of snag distribution maps. With the developed method, a map of the spatial distribution of spruce snags with a dbh greater than or equal to25 cm can be obtained in about 16 h for an area of 3 km(2) (i.e. 3 min/ha of map). Tree diameter, treetop condition (broken or intact) and the canopy closure of the forest stands significantly affected the success of snag detection. The method detected 82% (93%) of snags greater than or equal to25 cm (greater than or equal to35 cm) with an intact treetop, and 67% (71%) when broken snags greater than or equal to25 cm (greater than or equal to35 cm) were also included. Given our encouraging results, the method should be further tested on similar study sites in order to obtain more certainty regarding validity of the coefficient correcting underestimation. The method may become a promising tool, complementing standard field methods, with various prospective applications, such as wildlife studies, forest inventories, certification processes, etc. (C) 2004 Elsevier B.V. All rights reserved.