PALM-meteo: processor of meteorological input data for the PALM model system

PALM-meteo is an advanced and modular tool to create PALM's dynamic driver with initial and boundary conditions (IBC) and other time-varying data, typically using (but not limited to) outputs from mesoscale models. It is a successor to multiple older projects for PALM dynamic drivers, most importantly the deprecated WRF-interface project.

The documentation for PALM-meteo is available at https://palm-tools.github.io/palmmeteo/.

Functionality

The PALM-meteo workflow consists of these stages:

1. Model setup: setting up basic items such as the PALM model domain. Currently this requireds providing the already prepared PALM static driver.
2. Input loading: selection of requested variables, area selection, transformation and/or unit conversion where required. Certain input variables are also temporally disagregated.
3. Horizontal interpolation from the input model grid to PALM grid. Includes geographic projection conversion where required. Models with traditional rectangular grid are regridded using bilinear interpolation, the ICON model with the icosahedral grid uses Delaunay triangulation and barycentric interpolation.
4. Vertical interpolation from input model levels (which may be terrain-following, isobaric, eta, hybrid etc.) to PALM model levels (altitude-based). Part of this process is terrain matching, as the high-resolution PALM terrain may differ, even significantly, from the input model terrain. This process includes configurable vertical adaptation, where the lowest input layer is shifted to match the PALM terrain while in the higher layers, the vertical shifts are progressively smaller until they reach the transition level, above which they are not shifted at all.
5. Output generation creates the final PALM dynamic driver. Final adjustments are performed here, notably the mass balancing which is performed on all boundaries (respecting terrain on lateral boundaries), so that the PALM's internal mass balancing (which is performed only on the top boundary as a last resort) is not overused.

Currently PALM-meteo supports these meteorological inputs:

Meteorological IBC

• WRF
• ICON
• Aladin
• Synthetic inputs (detailed profile specification etc.)

Radiation inputs (optional)

• WRF
• ICON
• Aladin

Chemical IBC (optional)

• CAMx
• CAMS

PALM-meteo is higly modular and more input sources will be likely added in the future. A detailed technical description will be made available in the upcoming scientific paper.

Installation

There are three basic ways to install PALM-meteo.

Method 1: Simple minimal installation using PIP

PALM-meteo is available in PyPI and it can be installed with the simple command:

pip3 install palmmeteo

This will also install the pmeteo command for running PALM-meteo. Depending on your operating system and environment, you may want to create a virtual environment first, or use the --user option, or use pipx instead of pip.

However, this method will install only the bare minimum to run PALM-meteo, without documentation and tests, so it is only recommended for experienced users who want a quick installation.

Method 2: Full in-place installation with a virtual environment

This is the recommended method for most users.

1. Make sure that you have python3, pip, and python-venv installed and available. This is operating system dependent.
2. Download and extract PALM-meteo into a directory of your choice. If you have already installed the PALM model system, then PALM-meteo is already extracted under packages/dynamic_driver/palm_meteo, albeit probably not in the newest version. Or you can download PALM-meteo from the PALM gitlab server or from GitHub.
3. From within the PALM-meteo install directory, run ./setup. This will create a new virtual environment, install PALM-meteo with all dependencies and create a symlink to the pmeteo command (which enables the virtual environment automatically).

Method 3: Advanced installation for developers

1. Clone the PALM-meteo git repository.
2. Create and enable your own virtual environment if you prefer.
3. From within the PALM-meteo install directory, run ./setup_novenv. This will install PALM-meteo dependencies and create a simple pmeteo running script.

Testing

PALM-meteo comes with integration tests supplied. When installed using Method 2, the install script performs the tests at the end of a successful instalation. You may also exectute them at any time from the PALM-meteo install directory using the command

tests/integration_tests/all_tests.sh

or for indivudal tests:

tests/integration_tests/test_XX_NAMEOFTEST.sh

You may also examine the directory tests/integration_tests/simple_wrf as a reference WRF case for PALM-meteo.

Usage

For each dynamic driver, a YAML configuration file needs to be prepared (typically one per case, although it is possible to combine more files). This file uses sensible defaults for most options, so it does not need to be very long, as is demonstrated by the supplied file example.yaml. However for the beginners it is best to start by making a copy of the file template.yaml, which contains all user-selectable options with their defaults and documentation, and modifying it accordingly. See PALM-meteo configuration.

Basic model configuration

The main part of configuration is selecting a single or multiple tasks by adding a list item in the tasks: configuration section. Selecting a task means just telling what PALM-meteo what it has to do, which typically involves creating IBC and/or other PALM inputs using the selected method, such as using a specific input model.

These are the currently supported tasks (obviously many of them are mutually exclusive):

• wrf: Create IBC from WRF model outputs.
• wrf_rad: Create PALM radiation inputs from WRF model outputs (typically AUXHIST outputs with potentially different time step from standard WRFOUT).
• icon: Create IBC from ICON outputs in the NetCDF format.
• aladin: Create IBC from Aladin outputs in the grib format.
• camx: Create chemistry IBC from CAMx model outputs.
• cams: Create chemistry IBC from CAMS model outputs.

When the specified task(s) are selected, the task configuration mechanism enables the required plugins and pulls in the respective task-specific configuration defauls, which may be overwritten within the configuration file.

Running the model

With a prepared configuration file such as myconfig.yaml, simply run

./pmeteo -c myconfig.yaml

in the project directory. See Running PALM-meteo for more information.

Extending PALM-meteo

In order to add new input types or processing methods, PALM-meteo can be easily extended with user plugins. For documentation, see the Developer guide.

License and authors

PALM-meteo is distributed under the GNU GPL v3+ license (see the LICENSE file). It was created by the Institute of Computer Science of the Czech Academy of Sciences (ICS CAS) with contributions by the Charles University in Prague (MFF UK), the Czech Hydrometeorological Institute (CHMI) and the Deutsche Wetterdienst (DWD).

Acknowledgenments

PALM-meteo was created with the support of these projects:

• TURBAN: Turbulent-resolving urban modeling of air quality and thermal comfort (Technology Agency of the Czech Republic project TO01000219)
• CARMINE – Climate-Resilient Development Pathways in Metropolitan Regions of Europe (European Commission project 101137851)