Storing NetCDF data in a relational database

Climate science requires the analysis of variables such as temperature, precipitation, and ozone concentration, which depend on multiple dimensions, mainly time, latitude, longitude, and height/depth, and are commonly stored in the NetCDF file format. Various infrastructure has evolved around this format such as libraries, command line tools, visualization software, and web services. In this work we present a prototype for serving NetCDF data backed by the open-source RDBMS Postgres. We test various database schemas in the search for one that delivers response times low enough for visualization and further supports the diversity present among NetCDF files.