The Future of GLOSS Sea Level Data Archaeology

Abstract:

Long term climate records are rare, consisting of unique and unrepeatable measurements. However, data do exist in analogue form in archives, libraries and other repositories around the world. The Global Sea Level Observing System (GLOSS) Group of Experts aims to provide advice on locating hidden tide gauge data, scanning and digitising records and quality controlling the resulting data.

Long sea level data time series are used in Intergovernmental Panel on Climate Change (IPCC) assessment reports and climate studies, in oceanography to study changes in ocean currents, tides and storm surges, in geodesy to establish national datum and in geography and geology to monitor coastal land movement.

GLOSS has carried out a number of data archaeology activities over the past decade, which have mainly involved sending member organisations questionnaires on their repositories. The Group of Experts is now looking at future developments in sea level data archaeology and how new technologies coming on line could be used by member organisations to make data digitisation and transcription more efficient.

Analogue tide data comes in two forms

• charts, which record the continuous measurements made by an instrument, usually via a pen trace on paper
• ledgers containing written values of observations

The GLOSS data archaeology web pages will provide a list of software that member organisations have reported to be suitable for the automatic digitisation of tide gauge charts. Transcribing of ledgers has so far proved more labour intensive and is usually conducted by people entering numbers by hand. GLOSS is exploring using Citizen Science techniques, such as those employed by the Old Weather project, to improve the efficiency of transcribing ledgers. The Group of Experts is also looking at recent advances in Handwritten Text Recognition (HTR) technology, which mainly relies on patterns in the written word, but could be adapted to work with the patterns inherent in sea level data.