The Weather Research and Forecasting model (WRF) is a highly flexible model that can be used to investigate a wide range of atmospheric phenomena, like hurricanes or El Niño. The project would entail engaging with the model code, running some simulations (on a topic agreed with the supervisor), and processing their output.
WRF produces output in the Network Common Data Form format (NetCDF). This self-descriptive format contains its own metadata and is commonly used in atmospheric or oceanic simulations. NetCDF can be processed using Python, R, or a variety of other tools. The Geoscience Community Analysis Toolkit (GeoCAT) has implemented a set of examples in Python.
One of the benefits of WRF is that it features many different simulation options. There are publications comparing e.g. cloud and planetary boundary layer (PBL) schemes (Otkin and Greenwald 2008) and physics packages (Gallus and Bresch 2006). Settings for a given simulation are selected using a namelist.
Initially, students would run an ideal simulation to learn how the model works in its most basic form. Next, they would use the real version of the model for a test case simulation. Then they would select a particular aspect of microphysics, or a specific weather event, and design an experiment in collaboration with their supervisor to investigate their particular focus. This might involve comparing particular outputs like temperature or precipitation for several different model set-ups or against a set of observations. The WRF Preprocessing System (WPS) will be used to prepare input data for the model.
The model’s source code can be obtained from its GitHub repository, and there is an online tutorial that can give a sense of what is required to get the model running.
A good knowledge of UNIX and either Python or R is a prerequisite for this project. Students will need to install necessary libraries, compile and run the source code of the model, and analyse the output. The source code is in Fortran, which will probably be a new experience for most students. It is unlikely that students will actively engage with the source code, so the project will involve limited use of Fortran.
The project will focus on compiling, running, and analysing the model. Students will demonstrate their learning by simulating and processing data related to a particular type of microphysics, or a specific weather event. They will need to source and preprocess data for the model suitable for use in their particular simulations. The whole process will be documented in their report, with selected portions detailed in their oral and poster presentations.