Making Earth-System-Model (ESM) Data accessible is challenging due to the large amount of data that we are facing in this realm. The upload is time-consuming, expensive, technically complex, and every institution has their own procedures. Non-ESM experts face a lot of problems and pure data portals are hardly usable for inter- and trans-disciplinary communication of ESM data and findings, as this level of accessibility often requires specialized web or computing services. With the Model Data Explorer, we want to simplify the generation of web services from ESM data, and we provide a framework that allows us to make the raw model data accessible to non-ESM experts. Our decentralized framework implements the possibility for an efficient remote processing of distributed ESM data. Users interface with an intuitive map-based front-end to compute spatial or temporal aggregations, or select regions to download the data. The data generators (i.e. the scientist with access to the raw data) use a light-weight and secure python library based on the Data Analytics Software Framework (DASF, https://digital-earth.pages.geomar.de/dasf/dasf-messaging-python) to create a back-end module. This back-end module runs close to the data, e.g. on the HPC-resource where the data is stored. Upon request, the module generates and provides the required data for the users in the web front-end. Our approach is intended for scientists and scientific usage! We aim for a framework where web-based communication of model-driven data science can be maintained by the scientific community. The Model Data Explorer ensures fair reward for the scientific work and adherence to the FAIR principles without too much overhead and loss in scientific accuracy. The Model Data Explorer is in the progress of development at the Helmholtz-Zentrum Hereon, together with multiple scientific and data management partners in other German research centers. The full list of contributors is constantly updated and can be accessed at https://model-data-explorer.readthedocs.io.

<- Please cite both the article from preferred-citation and the upload on zenodo.