How to develop and apply a model data standard on microscale model data.

Numerical modeling makes it possible to represent complex processes in small-scale and complex areas like cities. For resolving obstacles, grid sizes in the order of meters are needed. Due to small grid sizes and numerical restrictions, such high-resolution investigations require a great deal of resources. Therefore, a re-use of the results by others would enhance the value of any of these model results. However, the subsequent use of model results is still poorly developed. Comparisons of model data, dissemination of results, or reproduction of simulations are hampered by inconsistent data structures, non-standardized variable names, and lack of information on model setup. In general, to ensure the reusability and accessibility of model data, data standards should be used. The most common data standard for atmospheric model output data are the CF conventions, a data standard for netCDF files, but this standard is currently not extended to cover the model output of obstacle resolving models (ORM).

The AtMoDat (Atmospheric Model Data) project developed a model data standard (ATMODAT standard) which ensure FAIR (Findability, Accessibility, Interoperability, and Reuse) and well documented data. We involved the micro-scale modelling community in this process with a web based survey (http://uhh.de/orm-survey) to find out which micro-scale ORMs are currently in use, their model specifics (e.g. used grid, coordinate system), and the handling of the model result data. Furthermore, the survey provides the opportunity to include suggestions and ideas, what we should consider in the development of the standard. We already identified typical variables used by ORMs (i.e. building structures, wall temperatures) and will propose them to be included in the CF convention.  The application of this standard is tested on the model output of the ORM MITRAS. The standard and experiences with its application will be presented.