EarthServer aims at open access and ad-hoc analytics on Earth Science (ES) data, based on the OGC geo service standards Web Coverage Service (WCS) and Web Coverage Processing Service (WCPS). The WCS model defines "coverages" as unifying paradigm for multi-dimensional raster data, point clouds, meshes, etc., thereby addressing most of Earth Science data. WCPS as aka "XQuery for raster data" allows declarative, SQL-style queries on coverages.The project will develop a pilot implementing these standards, integrated with NetCDF and GeoSciML, two core formats used in the ESs, and open-source GIS tools. Integration of WCPS with XQuery will allow mixed data/metadata queries. The unified service will support navigation, extraction, aggregation, and ad-hoc analysis on massive n-D ES data through queries of open-ended complexity, achieving a flexibility on coverage data as it is known from SQL. Clients will range from mobile devices over Web tools to high-end immersive virtual reality.- In the RTD part, EarthServer will establish open-source client and server technology which is scalable to Petabyte/Exabyte volumes, based on distributed processing, supercomputing, and cloud virtualization. Implementation will be based on the existing rasdaman server technology developed by one partner, an operational system similar to SciDB.- In the Service part, this new service will be installed on super-scale archives of data centers serving atmospheric, oceanography, geology, and general earth observation communities. Queries can span archive sites and cross-domain data sets. Volumes of 20+ TB will be demoed.- In the Networking part, user training, community involvement and outreach, education of society (students, pupils, etc.) and public authorities, and standardization will be performed (e.g., OGC and INSPIRE).This novel quality of service, with its unified overarching retrieval, will enable new discoveries and enhanced collaboration, both ad-hoc and in "long tail science".