SGS Database: use of relational databases to enhance data management for multi-site experiments

Abstract

Relational databases are increasingly used to manage large and complex experimental datasets. This technology was chosen to create a common database tool for a large multi-site experiment which measured aspects of the sustainability of various regionally focused grazing experiments, as part of the Sustainable Grazing Systems Program. Since the data needs of experiments expanded over time and the number of sites and subsites increased, it was necessary to adopt an evolutionary approach to creating the database. It was important that the database be customised to accommodate the wide range of experiments and the diverse needs of the many scientists involved and yet have complete conformity of structure to permit across-site queries. This paper describes the steps taken in this process and suggests how similar database projects could be implemented more efficiently in the future. As most of the scientists involved were initially new users of relational databases, there were some delays in achieving full adoption of a uniform approach to data handling. The software chosen was a commonly available desktop application which was modified using code developed to allow graphical queries, quality assurance of data, and interfacing with a purpose-built simulation model. A survey of users at the end of the project indicated that there was generally a high degree of satisfaction with the relational database developed, but it also identified areas requiring improvement. The need for developing mechanisms of sharing data and of protecting data for possible future use are briefly discussed.