Large　spatio-temporal　Earth　observation　(EO)　data　are　organized　as　multidimensional　data　cubes,　which　is　a　new　paradigm　for　users　to　interact　with　EO　data.　For　the　multidimensional　data　cubes,　geospatial　dimensions,　denoting　by　latitude　and　longitude,　are　indispensable.　However,　time　series,　climate　and　ocean　data,　etc.　also　require　time　axes　and　non-spatio-temporal　axes　in　the　n-dimensional　structure.　Therefore,　at　this　level,　a　major　challenge　in　data　cubes　interoperability　concerns　the　harmonization　of　domain　variables　among　different　data　cube　implementations.　Harmonization　of　data　cubes　from　data　with　significantly　different　spatio-temporal　reference　systems　always　requires　domain　variable　knowledge　base.　We　propose　to　identify　domain　variables　via　a　standardized　resolver　approach.　In　this　way,　cubes　interoperability　of　different　domain　variables　can　be　referred　to　and　interoperated　across　multidisciplinary　applications.　©　2021　IEEE.