Abhaas Goyal: metplan: Preprocessing package for the land models meteorological forcings
- Introduction and scientific context: Currently, Land Surface Models such as CABLE/JULES attempt to accept a wide range of CF-compliant datasets, internally translating or deriving the necessary variables required by the model. However, this increases technical debt, making it difficult to debug input parameter state; or maintain the codebase in general to adapt to different datasets. One of the proposals raised from the previous LWG Workshop was to have a consistent and standard set of input variables going into CABLE.
- Methodology/project summary: We introduce a prototype general-purpose preprocessor tool to map met-forcing data as input to Land Surface Models. Depending on the dataset, pre-processing stages include unit conversions, naming/attribution for the input data to be CF-compliant, and conversions of intermediate measures (such as calculating specific humidity from vapor pressure).
- Results: We create an easy-to-interpret configuration file for the user. The input dataset should contain all the standard variables required by the model - however, it can also provide optional variables if using another derivation formula. The resulting dataset would also follow the ALMA standards, the accepted standard for the community. Given the large volume and number of files, the tool would also parallelise these calculations wherever possible.
- Audience: Land Working Group, CABLE, JULES
- Keywords: Model development, Model Input, Data Processing