lagtraj Lagragian simulations trajectories

Producing a Lagrangian forcing

There are three steps to making forcing profiles with lagtraj:

1. Download source data domain for a given date-range (small for Eulerian simulations, big for Lagrangian)

2. Produce trajectory

3. Extract forcing profiles along the trajectory (with optional conversion to target a specific LES/GCM model)

The guide below first details how to install lagtraj and then guides your each of these three steps.

0. Getting started

Installing lagtraj

The most recent tagged version of lagtraj (and all its dependencies) can be installed with pip directly from pipy:

bash $> python -m pip install lagtraj

NOTE: if you are intending to modify lagtraj yourself you should check out the development notes.

lagtraj requires Python 3 and is developed and tested with python3.8 (in that we aim to follow the recommendations of NEP29) but later versions should work too (it may work with earlier versions too).

Once installed all lagtraj's commands are available from any directory and the follow the pattern

bash $> python -m lagtraj.<command>

lagtraj input and output

lagtraj stores everything (both source data and input definitions describing how domains, trajectories and forcings are set up) in a data directory (by default this will be data/ relative to where lagtraj is invoked). The directory structure is as follows:

bash data ├── domains │   ├── eurec4a_circle.yaml │   └── eurec4a_circle_data │   ├── an_model_2020-01-01.nc │   : │   └── fc_single_2020-01-03.nc ├── forcings │   ├── eure4a_20200103_lag_testcase.yaml │   └── eure4a_20200103_lag_testcase.nc └── trajectories    ├── eure4a_20200103_lag_testcase.yaml    └── eure4a_20200103_lag_testcase.nc

The name of each domain/trajectory/forcing inside lagtraj will be the full filename without the .yaml-extension. E.g. the domain definition in domains/eurec4a_circle.yaml will have the name eurec4a_circle inside lagtraj.

You can either make your own domain/forcing/trajectory definition (these are stored in yaml-files) by creating a yaml-file in the relevant directory or use one that lagtraj comes with. You can list the input-defintions bundled with your copy of lagtraj by running the following command:

bash $> python -m lagtraj.input_definitions.examples

Which will print

```bash The following domain/forcing/trajectory definitions are currently included with lagtraj:

lagtraj:// ├── domains │ ├── drydowncardingtonlocal │ ├── eurec4acircle │ └── eurec4anorthatlantic ├── forcings │ ├── drydowncardington2020042000eul │ ├── eurec4acampaigneulerian │ ├── eurec4a20200128first │ ├── eurec4a20200202firstshortpress │ ├── eurec4a20200202firstshort │ ├── eurec4a20200202first ... ```

1. Making source data available

lagtraj is based around making all data required for interpolation, integration and forcing calculation available before trajectories and forcings are calculated. This was done to reduce the number of data requests required to the data storage backends (e.g. ECMWF), but does mean that the expected spatial extent that a trajectory will reach must been known before performining a trajectory integration, otherwise lagtraj will issue a warning when the edge of the available domain is reached.

In order to download the ERA5 input data for lagtraj, you need an account with the Copernicus Data Store. You will also need to install the CDS api, see the api-howto.

Either create your own domain definition in data/domains/<domain_name>.yaml and run

bash $> python -m lagtraj.domain.download <domain_name> <start_date> <end_date>

Or use one of the domain definitions included with lagtraj (e.g. eurec4a_circle

bash $> python -m lagtraj.domain.download lagtraj://eurec4a_circle <start_date> <end_date> the <start_date> and <end_date> should be formatted as YYYY/MM/DD, e.g. 2020/02/02 for the 2nd of February 2020.

An optional flag --retry-rate <num_minutes> causes lagtraj to continue retrying download of submitted data requests every num_minutes minutes until all data has been downloaded. Every time this command is run it will attempt to download only data not yet downloaded.

You can monitor the status of your requests via the CDS requests page. Download times for model level data on the CDS can be somewhat variable.

2. Producing a trajectory

Once you have downloaded the required domain data you can either create a trajectory definition in data/trajectories/<trajectory_name>.yaml and run

bash $> python -m lagtraj.trajectory.create <trajectory_name>

Or use one of the trajectory definitions included with lagtraj (e.g. eurec4a_20200202_first_short

bash $> python -m lagtraj.trajectory.create lagtraj://eurec4a_20200202_first_short

The created trajectory will be stored in data/trajectories/<trajectory_name>.nc.

3. Producing forcing profiles

To produce forcings you need to create a forcing definition in data/forcings/<forcing_name>.yaml and run

bash $> python -m lagtraj.forcings.create <forcing_name> [--conversion <conversion_name>]

Or use one of the forcing definitions included with lagtraj (e.g. eurec4a_20200202_first_short)

bash $> python -m lagtraj.forcings.create lagtraj://eurec4a_20200202_first_short [--conversion <conversion_name>]

Forcing profiles conversion (targeting a specific GCM/LES)

When creating forcings it might be desirable to target a specific LES (Large-Eddy Simulation) model or GCM (Global Circulation Model) by converting the forcings to a specific format and setting parameters specific to the model being targeted. This can be achieved by using the --conversion flag and providing a conversion_name. lagtraj currently comes bundled with functionality to target the KPT LES and dephy LES format.

Conversion parameters are defined in a yaml-files similarly to how domain, trajectory and forcings definitions are stored, with one important difference: conversion definition files are associated with a specific forcing definition file (i.e. each forcing conversion definition points to only one specific forcing definition). lagtraj comes bundled with definitions for how to do forcing conversion with sensible defaults that you can modify for each forcing you wish to create.

To set the parameters for a conversion identifed by the name kpt for converting a forcing with name forcing_name you should create a file in data/forcings/<forcing_name>.<conversion_name>.yaml. Running a conversion will the convert data/forcings/<forcing_name>.nc and save to data/forcings/<forcing_name>.<conversion_name>.nc.

bash $> python -m lagtraj.forcings.create <forcing_name> [--conversion <conversion_name>]

Instead of creating a conversion definition starting from an empty file you can bootstrap the process by using the default parameters for a given target model included with lagtraj. This is achieved by using the lagtraj://-prefix when choosing the conversion name. E.g. to create the forcing named eurec4a_20200202_first_short bundled with lagtraj and have it converted to the dephy format with the default parameters you would run

bash $> python -m lagtraj.forcings.create lagtraj://eurec4a_20200202_first_short --conversion lagtraj://dephy

This will create the un-converted forcing in data/forcings/eurec4a_20200202_first_short.nc, the converted one in data/forcings/eurec4a_20200202_first_short.dephy.nc and the default conversion definition for targeting dephy will be copied to data/forcings/eurec4a_20200202_first_short.dephy.yaml. You can then modify the forcing parameters (for example change the number of levels) by editing data/forcings/eurec4a_20200202_first_short.dephy.yaml and rerunning the forcing creation with your local copy of the conversion definition (note the absence of the lagtraj:// prefix):

bash $> python -m lagtraj.forcings.create lagtraj://eurec4a_20200202_first_short --conversion dephy

You are of course welcome to rename the conversion however you like if for example you'd like to have multiple different converted version of the same forcings file.

Contributing and comments

If you have any comments/questions/issues about lagtraj please feel free to open an issue or have a look in docs/developing.md where we have added some notes on how to get started. All contributions are very welcome!