Relational databases for geoscience data: HYDROCHEM, a case study.

Millions of data points have been acquired or compiled through both phases of the Exploring for the Future (EFTF) program at Geoscience Australia (GA). This data that graces the EFTF Portal and appears in many publications has another home within specialist databases designed and built to house the specific data that GA collects. One such database is HYDROCHEM, which was implemented as part of the Enhanced Data Delivery (EDD) and National Groundwater Systems (NGS) projects. HYDROCHEM hosts 190,097 rows of groundwater, surface water and rainfall water chemistry analyses. This data was either previously hosted in the GNDWATER database, or compiled from legacy data stores. The redevelopment of GNDWATER to HYDROCHEM saw the de-duplication and updating of sample and site-specific metadata into other GA databases, such as SAMPLES, BOREHOLES and FIELDSITES. The redevelopment also added additional constraints to the database, including minimum metadata requirements, constrained look-up tables for units of measure, laboratory, method, filter sizes, standards and uncertainty types. Other features include minimum and maximum values for particular analytes and delivery of the data in standardised GA-preferred units of measure.