Abstract
UNAVCO and IRIS, major repositories for global geodetic and seismic
data, are in the process of joining their operations to form a unified
facility for supporting the broad spectrum of geophysical observations
and science required to help understand and predict the behavior of
Earth Systems. This process would be complicated in a static data
management environment, but both repositories are also migrating
archives and services to the cloud as part of the merger. To simplify
and unify archive data management, the organizations are collaborating
to create common data and metadata models for observations from a wide
variety of instruments and disciplines. For data, the initial focus has
been on the xArray data model, already used in the geodetic and
magnetotelluric communities, which can be implemented with several disc-
and cloud-native approaches (HDF5, netCDF4, and Zarr). For metadata, the
SensorML Standard developed by the Open Geospatial Consortium is being
explored because 1) SensorML accommodates the large parameter space
associated with instrument metadata required to use and trust complex
observations and 2) the ability to extend the standard when required.
The merger of two large repositories combined with migration to the
cloud requires careful identification and on-going testing of a wide
variety of assumptions about data management systems. This presentation
will focus on lessons learned so far.