A guide for digitising manuscript climate data

S. Brönnimann (1) , J. Annis (1) , W. Dann (1) , T. Ewen (1) , A. N. Grant (1) , T. Griesser (1) , S. Krähenmann (1) , C. Mohr (1) , M. Scherer (1) , C. Vogler (1)

1 IAC - Institute for Atmospheric and Climate Science [Zürich]

Hand-written or printed manuscript data are an important source for paleo-climatological studies, but bringing them into a suitable format can be a time consuming adventure with uncertain success. Before starting the digitising work, it is worthwhile spending a few thoughts on the characteristics of the data, the scientific requirements with respect to quality and coverage, and on the different digitising techniques. Here we briefly discuss the most important considerations and report our own experience. We describe different methods for digitising numeric or text data, i.e., optical character recognition (OCR), speech recognition, and key entry. Each technique has its advantages and disadvantages that may become important for certain applications. It is therefore crucial to thoroughly investigate beforehand the characteristics of the manuscript data, define the quality targets and develop validation strategies.

Domaines: Interfaces continentales, environnement; Sciences de la Terre

S. Brönnimann, J. Annis, W. Dann, T. Ewen, A. N. Grant, et al.. A guide for digitising manuscript climate data. Climate of the Past Discussions, 2006, 2 (3), pp.191-207. ⟨hal-00298127⟩