Quickstart Guide

This quickstart guide is for quick reference on how to download, build, and run GEOS-Chem Classic, which is the single-node instance of GEOS-Chem.

Tip

Please also see our GCHP Quickstart Guide if you would like to run GEOS-Chem across using more than one computational node.

This guide assumes that your environment satisfies GEOS-Chem Classic hardware and software requirements. This means you should load a compute environment such that programs like cmake are available before continuing. If you do not have some of the required software dependencies, you can find instructions for installing external dependencies in our Spack instructions.

For simplicity we will also refer to GEOS-Chem Classic as simply GEOS-Chem on this page. More detailed instructions on downloading, compiling, and running GEOS-Chem can be found in the User Guide elsewhere on this site.

1. Clone GEOS-Chem Classic

Download the source code:

$ git clone --recurse-submodules https://github.com/geoschem/GCClassic.git GCClassic
$ cd GCClassic

Tip

If you wish, you may choose a different name for the source code folder, e.g.

$ git clone --recurse-submodules https://github.com/geoschem/GCClassic.git my_code_dir
$ cd my_code_dir

Upon download you will have the most recently released version. You can check what this is by printing the last commit in the git log and scanning the output for tag.

$ git log -n 1

Tip

To use an older GEOS-Chem Classic version (e.g. 14.0.0), follow these additional steps:

$ git checkout tags/14.0.0                  # Points HEAD to the tag "14.0.0"
$ git branch version_14.0.0                 # Creates a new branch at tag "14.0.0"
$ git checkout version_14.0.0               # Checks out the version_14.0.0 branch
$ git submodule update --init --recursive   # Reverts submodules to the "14.0.0" tag

You can do this for any tag in the version history. For a list of all tags, type:

$ git tag

If you have any unsaved changes, make sure you commit those to a branch prior to updating versions.

2. Create a run directory

Important

The first time you create a run directory, you will be asked to provide registration information. Please answer all of the questions, as it will help us to keep track of GEOS-Chem usage worldwide. We will also add your information to the GEOS-Chem Users web page.

Navigate to the run/ subdirectory. To create a run directory, run the script ./createRunDir.sh:

$ cd run/
$ ./createRunDir.sh

Creating a run directory is interactive, meaning you will be asked multiple questions to set up the simulation. For example, running createRunDir.sh will prompt questions about configurable settings such as simulation type, grid resolution, meteorology source, and number of vertical levels. It will also ask you where you want to store your run directory and what you wish to name it, including whether you want to use the default name, e.g. gc_4x5_merra2_fullchem. We recommend storing run directories in a place that has a large storage capacity. It does not need to be in the same location as your source code. When creating a run directory you can quit and start from scratch at any time.

For demonstration purposes, we will use a full chemistry simulation run directory with the default name (gc_merra2_4x5_fullchem). The steps to setup and run other types of GEOS-Chem Classic simulations follow the same pattern as the examples shown below.

Attention

The convection scheme used to generate archived GEOS-FP meteorology files changed from RAS to Grell-Freitas starting 01 June 2020 with impact on vertical transport. Discussion and analysis of the impact is available at https://github.com/geoschem/geos-chem/issues/1409.

To fix this issue, different GEOS-Chem convection schemes are called based on simulation start time. This ensures comparability in GEOS-Chem runs using GEOS-FP fields generated using the RAS convection scheme and fields generated using Grell-Freitas, but only if the simulation does not cross the 01 June 2020 boundary. We therefore recommend splitting up GEOS-FP runs in time such that a single simulation does not span this date. For example, configure one run to end on 01 June 2020 and then use its output restart to start another run on 01 June 2020.. Alternatively consider using MERRA2 which was entirely generated with RAS, or GEOS-IT which was entirely generated with Grell-Freitas. If you wish to use a GEOS-FP meteorology year different from your simulation year please create a GEOS-Chem GitHub issue for assistance to avoid accidentally using zero convective precipitation flux.

3. Load your environment

Always make sure that all libraries and environment variables are loaded prior to building GEOS-Chem Classic. An easy way to do this is to write an environment file and load that file every time you work with GEOS-Chem. To make this extra easy you can create a symbolic link to your environment file within your run directory or for reference. For example, do the following in your new run directory to have a handy link to the environment you plan on using.

$ cd /path/to/gc_4x5_merra2_fullchem   # Skip if you are already here
$ ln -s ~/envs/gcc.gfortran10.env gcc.env

Then every time you start up a session to work with GEOS-Chem in your run directory you can easily load your environment.

$ source gcc.env

4. Configure your build

You may build GEOS-Chem Classic from within the run directory or from anywhere else on your system. But we recommend that you always build GEOS-Chem Classic from within the run directory. This is convenient because it keeps all build files in close proximity to where you will run the model. For this purpose the GEOS-Chem run directory includes a build directory called build/.

First, navigate to the build/ folder of your run directory:

$ cd /path/to/gc_4x5_merra2_fullchem  # Skip if you are already here
$ cd build

The next step is to configure your build. These are persistent settings that are saved to your build directory. A useful configuration option is -DRUNDIR. This option lets you specify one or more run directories that GEOS-Chem is “installed” to; that is, where where the executable is copied, when you do make install.

Configure your build so it installs GEOS-Chem to the run directory you created in Step 2. The run directory is one directory level higher than the build directory. Also located one level higher than the build directory is the CodeDir symbolic link to the top-level GEOS-Chem source code directory. Use the following command to configure your build:

$ cmake ../CodeDir -DRUNDIR=..

GEOS-Chem has a number of additional configuration options you can add here. For example, to compile with RRTMG after running the above command:

Note

The . in the cmake command above is important. It tells CMake that your current working directory (i.e., .) is your build directory.

$ cmake . -DRRTMG=y

A useful configuration option is to build in debug mode. Doing this is a good idea if you encountered an error (such as a segmentation fault) in a previous run and need more information about where the error happened and why.

$ cmake . -DCMAKE_BUILD_TYPE=Debug

Click here for more information on configuration options.

5. Compile and install

Compiling GEOS-Chem Classic should take about a minute, but it can vary depending on your system, your compiler, and your configuration options. To maximize build speed you should compile GEOS-Chem in parallel using as many cores as are available. Do this with the -j flag from the build/ directory:

$ cd /path/to/gc_4x5_merra2_fullchem/build   # Skip if you are already here
$ make -j

Upon successful compilation, install the compiled executable to your run directory:

$ make install

This copies executable build/bin/gcclassic and supplemental files to your run directory.

Note

You can update build settings at any time:

1. Navigate to your build directory.

2. Update your build settings with cmake (only if they differ since your last execution of cmake)

3. Recompile with make -j. Note that the build system automatically figures out what (if any) files need to be recompiled.

4. Install the rebuilt executable with make install.

If you do not install the executable to your run directory you can always get the executable from the directory build/bin.

6. Configure your run directory

Now, navigate to your run directory:

$ cd /path/to/gcc_4x5_merra2_fullchem

You should review these files before starting a simulation:

Configuration file	Description
geoschem_config.yml	Controls several frequently-updated simulation settings (e.g. start and end time, which operations to turn on/off, etc.)
HISTORY.rc	Controls GEOS-Chem diagnostic settings.
HEMCO_Diagn.rc	Controls emissions diagnostic settings via HEMCO.
HEMCO_Config.rc	Controls which emissions inventories and other non-emissions data will be read from disk (via HEMCO).

Attention

If you wish to spin up a GEOS-Chem simulation with a restart file that has (1) missing species or (2) a timestamp that does not match the start date in geoschem_config.yml, simply change the time cycle flag for the SPC_ entry in HEMCO_Config.rc from

* SPC_ ... $YYYY/$MM/$DD/$HH EFYO xyz 1 * - 1 1

to

* SPC_ ... $YYYY/$MM/$DD/$HH CYS xyz 1 * - 1 1

This will direct HEMCO to read the closest date available (C), to use the simulation year (Y), and to skip any species (S) not found in the restart file.

Skipped species will be assigned the initial concentration (units: \(mol\ mol^{-1}\) w/r/t dry air) specified by its BackgroundVV entry in species_database.yml. If the species does not have a BackgroundVV value specified, then its initial concentration will be set to \(1.0{\times}10^{-20}\) instead.

Please see our Customize simulations with research options Supplemental Guide to learn how you can customize your simulation by activating alternate science options in your simulations.

7. Download input data

Before you can run your GEOS-Chem Classic simulation, you must first download the required input data. These data include:

• Meteorological fields (e.g. GEOS-FP, MERRA-2, GEOS-IT, or GCAP2)

• Emissions inventories

• Inputs for GEOS-Chem modules (e.g. Cloud-J)

• Initial conditions for starting GEOS-Chem simulations

Tip

If your institution has several GEOS-Chem users, then someone may have already downloaded these data for you. If this is the case, you may start running your your GEOS-Chem Classic simulation right away.

7a. Run GEOS-Chem Classic in “dry-run” mode

The easiest way to download data is to perform a dry-run simulation. This is a GEOS-Chem Classic simulation that steps through time, but does not perform computations or read data files from disk. Instead, the dry-run simulation prints a list of all data files that the simulation would have read.

To start a dry-run simulation, type this command:

$ ./gcclassic --dryrun | tee log.dryrun

This will generate the log.dryrun log file, which contains the list of data files to be downloaded.

7b. Download the data (requires Python)

Once the dry-run simulation has finished, use the download_data.py file (included in your run directory) to download the required data.

You will need to activate a Python environment before you can download the data. We recommend using the Python environment for GCPy, as it has all of the relevant packages installed. If you have installed GCPy from PyPI, then no further action is needed. On the other hand, if you installed GCPy from conda-forge, you will need to activate the GCPy Python environment with this command:

$ conda activate gcpy_env

The prefix (gcpy_env) will be added to the command line prompt, which lets you know that the Python environment is active. (If you installed GCPy from PyPI, you will not see this prefix.)

You may now begin downloading data from the GEOS-Chem Input Data portal:

(gcpy_env) $ ./download_data.py log.dryrun geoschem+http

The data will be transfered using HTTP protocol. But if you have AWS CLI (command line interface) installed on your system, you can this command instead:

(gcpy_env) $ ./download_data.py log.dryrun geoschem+aws

This will use the AWS CLI data download protocol instead, which should be faster than regular HTTP connections. This is the command you should use if you are running GEOS-Chem Classic in an AWS EC2 instance.

If you have activated the GCPy Python environment, you may now deactivate it:

(gcpy_env) $ conda deactivate
$

This will remove the (gcpy_env) prefix from the command prompt.

At this point the required data files for your simulation should have been successfully downloaded from the GEOS-Chem Input Data portal to your computer system or EC2 instance.

7c. (Optional) Download additional meteorology data

You may need to perform a subsequent dry-run simulation to download additional data that are stored separately from the GEOS-Chem Input Data portal:

1. If you plan to run a GEOS-Chem Classic nested-grid simulation with meteorology fields that have been cropped to a specific nested grid domain, then follow these steps:

   $ ./gcclassic --dryrun | tee log.dryrun.nested
   
   $ conda activate gcpy_env                                      # Skip if using GCPy from PyPI
   
   (gcpy_env) $ ./download_data.py log.dryrun.nested nested+http  # or nested+aws if you have AWSCLI
   
   $ conda deactivate                                             # Skip if using GCPy from PyPI
   

   This will download the cropped meteorology fields from our GEOS-Chem Nested Input Data portal to your computer system or EC2 instance.

2. If you plan to perform a GEOS-Chem Classic simulation drven by GCAP 2.0 meteorology, follow these steps:

   $ ./gcclassic --dryrun | tee log.dryrun.gcap2
   
   $ conda activate gcpy_env                                      # Skip if using GCPy from PyPI
   
   (gcpy_env) $ ./download_data.py log.dryrun.gcap2 rochester
   
   $ conda deactivate                                             # Skip if using GCPy from PyPI
   

   This will download the GCAP 2.0 meteorology data from the GCAP 2.0 data portal hosted at U. Rochester to your computer system or EC2 instance.

For more information about dry-run simulations, please see our Download data with a dry-run simulation chapter.

8. Run your simulation

If you used an environment file to load software libraries prior to building GEOS-Chem then you should load that file prior to running. To run GEOS-Chem Classic, type at the command line:

$ ./gcclassic

If you wish to send output to a log file, use:

$ ./gcclassic > GC.log 2>&1

We recommend running GEOS-Chem Classic as a batch job, although you can also do short runs interactively. Running GEOS-Chem as a batch job means that you write a script (usually bash) and then you submit that script to your local job scheduler (SLURM, LSF, etc.). If you write a batch script you can include sourcing your environment file within the script to ensure you always use the intended environment. Submitting GEOS-Chem as a batch job is slightly different depending on your scheduler. If you aren’t familiar with scheduling jobs on your system, ask your system administrator for guidance.

Those are the basics of using GEOS-Chem Classic! See this user guide, step-by-step guides, and reference pages for more detailed instructions.